Executive takeaways

A tariff can change the economics of a sourcing decision quickly. Moving production, qualifying a new supplier or building a factory requires a commitment that may last far longer than the policy that helped justify it, said David Jacoby in a presentation at the recent ASCM CHAINge conference in Long Beach, California.

Jacoby, a supply chain consultant who teaches at New York University and the University of Pennsylvania, outlined a range of possible tariff paths moving forward and urged companies to examine how their operations would perform under different outcomes.

His message for executives was that uncertainty still leaves room for action. Companies can evaluate alternative production locations, model disruptions and structure supplier relationships to share risk. The challenge is deciding which changes remain useful if the tariff assumptions change.

“So if you’re managing global operations, there are a lot of potential strategies—that’s the good news—you can do a lot actually,” he said.

Jacoby’s outlook included a gradual reduction in tariffs over an extended period, rather than an immediate return to a low-tariff environment. He presented that as his judgment, but for companies considering substantial capital commitments, the range of possibilities matters as much as his preferred scenario.

Match the investment horizon to the policy risk

Jacoby argued that uncertainty about the durability of trade policy can make companies reluctant to commit to major investments. A business evaluating a new manufacturing location needs to consider whether the conditions supporting the project will persist long enough to justify its cost.

In his view, U.S. election cycles complicate that calculation. He contrasted a long period of policy certainty, which companies could incorporate into capital decisions, with the possibility that the policy direction changes within a few years.

The implication for a supply chain leader is to examine the investment across more than one tariff assumption. A project that appears attractive under high tariffs may produce a different return if duties fall. Conversely, an alternative sourcing arrangement may have value beyond tariff avoidance if it improves service continuity or reduces dependence on a single location.

Jacoby described several broad paths, including sustained high tariffs, lower tariffs, an intermediate level and a phased reduction. He considered a gradual decline over 10 or 20 years the most likely longer-term direction, while acknowledging that events could prove him wrong.

That qualification is important to the investment discussion, he said. His outlook provides a scenario to test, rather than a policy timetable on which companies can safely base a business case.

For executives, the useful question is how a proposed investment performs if the forecast is wrong. That means examining both the benefits of a change and the commitments it creates, including the cost and time required to change course again.

Evaluate where each stage of production belongs

Jacoby discussed diversification into other manufacturing locations as one response to tariff exposure. He also cautioned that relocating production to the United States involves considerations beyond the duty applied at the border.

Drawing on conversations with foreign companies at conferences and trade shows, he said he had encountered resistance to establishing U.S. manufacturing operations. He cited labor costs and the business relationship with the United States among the concerns he heard.

For U.S. companies already producing abroad, he said reshoring can make sense when the operation is structured appropriately. In response to an audience question, he pointed to combinations of nearshoring, partial assembly and final assembly in the United States.

A change in location must still provide the necessary production capability, workforce, supplier support and customer service, he said. Lower tariff exposure is just one part of the business case.

Make supplier contracts part of the response

One of Jacoby’s clearest recommendations concerned supplier relationships. He urged companies to structure contracts around the allocation of risk, alongside the price of the product or service.

As Jacoby put it, “contracting is about risk sharing, not just price.”

That approach broadens the procurement conversation. When trade conditions change, a supplier and customer need to understand how the change affects their relationship and which responsibilities each party has assumed.

Jacoby said there are multiple ways to structure contracts for risk sharing and encouraged companies to examine those options alongside network design. He did not present a single contractual solution that would work for every business.

For executives, the starting point is to identify the exposure the relationship is intended to manage. That review can inform discussions with suppliers about how the arrangement should work under the scenarios the company is testing.

The objective is to make the commercial relationship part of the operating response. A network strategy that depends on suppliers behaving in a particular way needs agreements and expectations that support it.

Test recovery as well as cost

Jacoby also recommended running scenarios that measure the time required to recover from a disruption and its expected effect on performance. His examples extended beyond tariffs to other events that can interrupt a global operation.

That wider lens matters because a sourcing change made to address trade policy can alter other risks. Companies need to understand how a proposed network functions when an interruption occurs, as well as how much it costs under normal conditions.

Recovery time provides one way to assess whether an alternative can become operational when needed. Expected performance impact helps connect the disruption to consequences for the business. Together, those measures can make the resilience discussion more concrete than simply counting additional suppliers or locations.

Jacoby linked that analysis to the need for appropriate metrics and systems that support agility. A company’s data and operating platforms must allow it to evaluate changes and shift production when the strategy requires it, he said.

The capability to act therefore belongs in the scenario itself. Identifying an alternative location does not establish how quickly production can move there or whether the company can manage the transition effectively.

Bring the choices into the governance discussion

Jacoby closed his formal presentation by connecting supply chain risk with board oversight. He referenced governance practices he had developed for a presentation to the National Association of Corporate Directors and encouraged attendees to consider how their boards govern both general risk and supply chain exposure.

That places tariff uncertainty within a broader management discussion. Sourcing changes, capital investments and supplier commitments can affect the business well beyond the supply chain function. The assumptions supporting those choices need to be understood by the people approving them.

Jacoby’s presentation offered a forecast, but its more durable contribution was a set of questions executives can apply across different outcomes: How does the investment perform if tariffs change? Which production stages can move? How is supplier risk shared? How long would recovery take?

Those questions give companies a way to keep making decisions while the policy outlook remains unsettled.

Today is Tuesday, Oct. 20

1. Japan’s yttrium supply warning exposes a critical-minerals vulnerability

Japanese officials warn domestic yttrium supplies could run out within months as Chinese export restrictions tighten access to a rare earth used in vehicles, jet engines and semiconductors.

Financial Times | Read the original article

2. Chinese technology groups place multibillion-dollar Nvidia orders

ByteDance, Alibaba and Tencent reportedly ordered more than three million export-compliant Nvidia workstation chips worth over $35 billion, but production constraints mean the manufacturer cannot guarantee requested volumes.

Reuters | Read the original article

3. Asian refiners redesign crude sourcing around tanker constraints

Asian refiners are shifting toward smaller tankers, long-term charters and delivered Middle Eastern crude as near-record freight rates make longer-haul oil from the United States and Africa less competitive.

Reuters |  Read the original article

4. Trans-Pacific reliability decline raises booking risk

Asia-to-North America on-time performance fell to 26.9% in September as port disruption and reduced planned capacity increased booking risk, prompting shippers to reserve space earlier and add delivery buffers.

FreightWaves | Read the original article

5. Broker-liability verdict heightens carrier-vetting concerns

A Texas judge affirmed a $604 million verdict involving C.H. Robinson, heightening broker concerns about carrier vetting, vicarious liability and the standards courts apply when defining operational control.

Logistics Management |  Read the original article

6. Trucking recovery depends on capacity exiting the market

ATA’s chief economist said trucking’s recovery is being driven by shrinking capacity rather than stronger demand, after years of falling revenue per mile and rising nonfuel operating costs forced carriers out.

Transport Topics |  Read the original article

7. Fuel costs lift truckload rates while demand remains subdued

September truckload rates rose across all equipment types as diesel surpassed $6 per gallon, while underlying volumes remained below year-earlier levels and fuel-adjusted van and flatbed rates declined.

Logistics Management |  Read the original article

8. Walmart broadens parcel options for marketplace sellers

Walmart Marketplace added eight parcel carriers for seller-fulfilled orders, broadening delivery choices beyond the largest national providers while retaining tracking requirements that can affect seller account standing.

Supply Chain Dive |  Read the original article

9. AI becomes a strategic tool for omnichannel fulfillment

MIT researchers describe AI as a strategic enabler for omnichannel fulfillment, examining how supply chain teams use the technology to manage operational complexity and address persistent execution challenges.
DC Velocity |  Read the original article

10. Parcel strategies evolve around capacity, service and AI

AlixPartners’ Marc Iampieri examines how parcel diversification, peak-season capacity, rising delivery expectations and AI-assisted routing are reshaping logistics strategies as retailers balance service performance against higher final-mile costs.

Logistics Management |  Read the original article

Executive takeaways

For the past several years, supply chain leaders have focused on a fundamental question: How quickly can artificial intelligence transform planning, sourcing, inventory management, logistics, and operations?

That question is rapidly becoming less important.

The latest research from ASCM and the Digital Supply Chain Institute suggests that executive skepticism is no longer the primary barrier to AI adoption. Most senior leaders now recognize AI's strategic importance. The more pressing challenge is organizational: Can the enterprise adapt quickly enough to capture the value the technology makes possible?

(The full research can be downloaded here)

Across industries, the gap is no longer between AI capabilities and executive vision. It is between AI capabilities and organizational capacity.

The companies that create competitive advantage in the next phase of AI adoption will not necessarily be those with the most advanced algorithms. They will be the organizations that most effectively redesign work, decision-making, talent models, governance structures, and cross-functional collaboration to leverage AI at scale.

In other words, AI transformation is becoming an organizational-design challenge as much as a technology challenge.

The technology is advancing faster than the enterprise

The research reveals an interesting paradox. AI activity is increasing across supply chain functions, yet true transformation remains relatively limited.

Most organizations report using AI primarily to improve analysis, forecasting, visibility, and recommendations. Measured on a 1-to-5 scale, where 1 means no AI use and 5 means fully autonomous execution without human review, the average adoption score across 13 core supply chain functions is just 1.94. Actual decision autonomy remains low. Even in demand planning, the most advanced function at 2.28, 35% of organizations report no AI use at all, and only a small fraction allow AI to execute decisions without review. In returns and reverse logistics, the least advanced function at 1.59, 73% report no AI use. The average adoption score across functions suggests that AI is still largely being used to assist decision-makers rather than fundamentally change how decisions are made.

 

This distinction matters.

For many organizations, AI implementation has focused on improving existing processes. The next wave of value creation will come from redesigning processes altogether. Leaders must begin asking not only where AI can improve performance, but how work itself should be structured in a world where intelligence is abundant, scalable, and increasingly automated.

That requires a different conversation. It is less about technology deployment and more about organizational architecture.

The most important finding may be the middle-management gap

Among the study’s most revealing findings is the disconnect between executive enthusiasm and managerial engagement.

C-suite leaders are actively involved in AI initiatives at remarkably high rates. Fully 82% of C-suite leaders report that they lead, implement, or scale AI initiatives, and vice presidents follow closely at 72%. Yet director-level engagement drops sharply. Only 12% of directors directly lead AI efforts, and half (50%) describe themselves as merely aware of them. Among managers, 51% sit in the aware-only category, a figure that rises to 70% for staff and associates, with another 20% of staff reporting no involvement at all. Many directors and managers remain observers rather than active leaders of AI transformation.

 

This gap represents one of the greatest risks to enterprise AI success.

Transformation does not occur in boardrooms. It occurs where workflows are redesigned, operational tradeoffs are made, teams are managed, and performance is measured. Directors and managers serve as the critical translation layer between executive ambition and operational reality.

If executives are setting AI strategy but middle management has not been empowered, trained, or incentivized to lead change, AI initiatives are likely to stall. Many companies appear to have AI strategies. Fewer have AI operating models. The leaders who close this middle-management gap may ultimately outperform organizations that invest far more heavily in technology.

AI struggles most where organizations struggle most

Another important insight emerges from the planning functions.

Demand planning and inventory planning show stronger adoption rates than sales & operations planning (S&OP). On the same 1-to-5 scale, demand planning scores 2.28 and supply and inventory planning scores 2.12, while S&OP scores just 1.93, trailing every other planning category, with 44% of organizations reporting no AI use in S&OP at all. On the surface, this appears surprising. S&OP is one of the most data-intensive and analytically dependent business processes in the enterprise.

But S&OP introduces a challenge that forecasting functions often do not: organizational complexity.

Demand planning can largely operate within a defined function. S&OP requires alignment among sales, operations, finance, procurement, manufacturing, and executive leadership. It crosses organizational boundaries, competing incentives, and multiple sources of authority.

The finding suggests that AI adoption becomes more difficult when technology must navigate organizational seams rather than technical challenges. This may be a lesson that extends beyond supply chain management.

Many organizations have spent years investing in digital transformation. AI may now be revealing a deeper issue: technology can move quickly, but organizations often struggle to move across functional boundaries. As AI becomes more capable, organizational friction may increasingly replace technical limitations as the primary constraint on performance.

Trust and decision rights are the next frontier

The research also highlights an important reality: most organizations remain cautious about delegating authority to AI systems.

That caution is understandable. Supply chains operate in environments characterized by uncertainty, disruption, and risk. Leaders want explainability, transparency, and accountability before allowing automated systems to influence critical decisions.

However, the challenge is not simply one of trust. It is one of decision design.

The most important question facing executives over the next several years may be: Who decides what?

Organizations will increasingly need formal frameworks that define: decisions informed or recommended by AI, decisions executed by AI with human oversight, and decisions executed autonomously within established guardrails.

This is not merely a technology issue. It is a governance issue. The organizations that solve this challenge first will likely gain advantages in speed, consistency, scalability, and resilience.

Readiness comes through action

One of the most encouraging findings from the research is the strong relationship between organizational readiness and project success. Organizations that view themselves as AI-ready report substantially higher success rates than those that consider themselves unprepared. The 17% of organizations with an AI-ready workforce report a 71% project success rate, compared with 28% for the 83% that consider their workforce unprepared. The prepared group has put AI to work across an average of six supply chain functions, while roughly 64% of the unprepared group remains stuck in pilot or planning stages.

Executives should be careful not to assume causation. However, the relationship reinforces an important lesson: readiness is often built through implementation rather than before it.

Many organizations continue waiting for perfect data, perfect governance, perfect training, or perfect alignment before scaling AI initiatives. That moment is unlikely to arrive.

Leading organizations are learning by doing. They are building governance while deploying use cases. They are increasing workforce capability through experience. They are refining standards through real-world implementation. The path to AI maturity may be iterative rather than sequential.

Five actions for supply chain leaders

Executives looking to accelerate AI transformation should consider several priorities.

1. Focus on organizational design, not just technology. AI investment strategies should be accompanied by redesign efforts that address workflows, processes, accountability, incentives, and decision-making structures.
2. Turn directors and managers into AI leaders. Middle managers must become active participants in transformation efforts rather than passive recipients of executive mandates. Training, authority, and incentives should reflect this responsibility.
3. Redesign decision rights. Every organization should develop a structured framework defining where AI informs, recommends, executes, or autonomously acts.
4. Prioritize cross-functional use cases. S&OP, integrated business planning, supplier collaboration, and end-to-end visibility initiatives often generate greater strategic value because they address organizational coordination challenges rather than isolated functional efficiencies.
5. Build trust through governance. Explainability, transparency, validation, and human accountability should be treated as strategic enablers of adoption rather than compliance requirements.

The next competitive advantage

For supply chain leaders, the central question is no longer whether AI will transform operations. It will. The more important question is whether organizations can transform themselves quickly enough to take advantage of it.

The ASCM-DSCI research suggests that workforce readiness, middle-management engagement, organizational alignment, governance, and trust are increasingly the factors that separate AI leaders from AI followers. Technology remains critical, but technology alone is not sufficient.

The companies that thrive in the AI era may not be those with the most advanced models. They may be those that most effectively redesign their organizations to work alongside them.

About the author

Vivek Ghelani is a global leader in supply chain transformation and digital strategy. As Director of Research at the Digital Supply Chain Institute (DSCI), an applied research institute of the Center for Global Enterprise in New York, he drives research and executive engagement that help organizations transform traditional supply chains into intelligent, connected, and customer-centric networks of value.

Executive takeaways

One of the largest supply chain costs in a hospital may never appear as a line item. It shows up in the extra work created when supply chain processes don’t work as they should, whether a nurse needs to search for a product or an invoice requires a manual correction. Across a health system, those extra minutes can carry a substantial cost.

At a hospital with 100 nurses, losing just 10 minutes per nurse each day searching for supplies can consume thousands of clinical hours over a year. The labor cost can reach hundreds of thousands of dollars. That is time clinicians are spending dealing with a supply chain issue instead of supporting patient care.

Fragmented systems and poor process design create this type of operational friction. While most health systems track purchase price savings closely, few account for the time consumed by workarounds and slow decisions.

Friction hides inside familiar work

Friction often survives because employees build workarounds that become part of the routine. When a clinician cannot locate an item through the usual system, an unofficial spreadsheet or a call to another location may keep the work moving. The workaround solves the immediate problem while allowing the underlying process weakness to remain.

To address the challenge, leaders can start by identifying routine transactions that require employees to step in. From there, they can determine if a workflow redesign or consolidation of systems can reduce bottlenecks and simplify process handoffs.

One health system addressed friction by bringing its hospitals onto a common enterprise resource system and redesigning purchasing and invoicing processes that required frequent manual intervention. Examining where routine transactions unnecessarily require manual work can show leaders where processes are consuming unnecessary staff time and help identify opportunities to streamline.

Automation exposes weak processes

Pressure to modernize can tempt organizations to buy advanced tools while core workflows still depend on manual intervention. Instead of offering an improvement, technology placed over a fragmented process usually preserves the fragmentation and makes failures harder to diagnose.

Automation is most useful when the organization understands why employees intervene in a workflow. In healthcare, that means distinguishing unnecessary administrative variation from differences required by clinical practice. Technology can then absorb routine work without reproducing the handoffs responsible for the delay.

One health system found that processing surgical implant purchases involved considerable repetitive work. After simplifying the process, the organization automated much of that work, saving thousands of employee hours and hundreds of thousands of dollars in the first year. The results show why the process must be addressed before automation is introduced.

AI should make decisions easier

AI can also reduce friction in health system work that requires employees to review large amounts of information before they can act. Contracting is one example. At one health system, AI reduced the time employees spent extracting information and reviewing routine contract language. Contract turnaround fell by more than half, giving the existing team more capacity.

The same idea extends to supply risks. AI can help teams spot potential shortages earlier, giving employees more time to determine how to respond. That additional visibility helped one health system significantly reduce backorders after it combined internal inventory information with supplier data.

Seeing a risk sooner only helps if someone can act on it. When ownership is unclear or a decision must move through multiple approvals, better information can still sit idle. Health systems need to examine the work surrounding AI. Leaders need to understand who receives the information, who can make the decision and how much manual effort stands between the two. That may mean clarifying decision authority or eliminating approvals that add time without changing the outcome.

Make friction measurable

Purchase price does not account for the clinical time consumed when supplies are difficult to locate or decisions stall during a disruption. Measuring that lost time can help a health system determine where process redesign deserves attention and whether technology is reducing the burden.

Healthcare supply chain leaders should apply the same scrutiny to time that they already apply to cost. When a process is redesigned or new technology is introduced, the question should be whether employees are spending less time working around the supply chain and can respond more quickly when something changes.

For health systems under sustained financial pressure, time recovered from friction is a cost worth putting on the scorecard.

Gartner will announce the Gartner Healthcare Supply Chain Top 25 for 2026 in a webinar on November 18.

About the author

Bruce Gilmore is a VP Analyst in Gartner's Supply Chain Practice focused on healthcare provider supply chain and procurement strategy.

Today is Monday, Oct. 19

1. Commercial-vessel attacks intensify in the Strait of Hormuz

At least 10 commercial vessels were attacked near the Strait of Hormuz in one week, including a Singapore-flagged container ship that caught fire and sank off Oman.

FreightWaves  |   Read the original article

2. Russian-fuel waiver adds another variable to energy planning

Oil retreated from five-month highs after the United States waived key Russian sanctions and tariffs, even as tanker attacks reduced Gulf crude flows to roughly 8.5 million barrels daily.

Reuters  |  Read the original article

3. Strategic-industry funding reaches another $5 billion

JPMorgan said it has invested $5 billion in companies supporting advanced manufacturing, energy, supply chains and defense technologies, alongside about $230 billion financed through its resiliency initiative.

Reuters  |   Read the original article

4. China’s export engine masks persistent demand weakness

China’s economy grew 4.5% in the third quarter as exports and industrial output offset weak domestic consumption, increasing exposure to tariffs, protectionist measures and geopolitical friction.

Reuters  |  Read the original article

5. Long Beach cargo volume approaches one million containers

The Port of Long Beach handled a record 997,177 TEUs in September, up 25% year over year, as loaded imports rose sharply without reported terminal congestion.

FreightWaves  |   Read the original article

6. Cross-border carriers warn enforcement could constrain capacity

U.S. and Mexican transportation leaders said inconsistent enforcement of cabotage, visa, English-language and driver-qualification rules could delay compliant shipments and complicate driver scheduling across the southern border.

FreightWaves  |  Read the original article

7. Sustainability consensus runs ahead of logistics execution

A 1,085-person global survey found broad agreement on logistics decarbonization measures, yet infrastructure and feasibility remain barriers; 83% expect climate adaptation to force network redesign, but 45% call current efforts adequate.

DC Velocity  |   Read the original article

8. Sustainability software shifts toward operating decisions

An industry analysis argues that emissions and ESG systems create more value when auditable data informs sourcing, transportation, facility and product choices instead of remaining confined to compliance reporting.

Logistics Viewpoints  |   Read the original article

9. Product records become a resilience control point

Manufacturers can shorten disruption response by maintaining approved alternate parts, supplier qualifications, revision effectivity and engineering-change propagation in a governed product record connected to procurement and operations.

Supply Chain Management Review  |   Read the original article

10. Supply-chain redesign must account for workforce effects

Researchers urge leaders to treat worker well-being as a design constraint, measuring labor volatility, safety exposure and downstream workforce impacts alongside cost and service when deploying AI and automation.

Supply Chain Management Review  |   Read the original article

Executive takeaways

A crisis playbook can tell a supply chain team where to start. It cannot anticipate every supplier failure, staffing constraint or sourcing decision that could interrupt the business. The harder work is building an organization that can recognize those vulnerabilities and adapt before a disruption becomes an emergency.

That was the central theme of a panel discussion from the recent ASCM CHAINge conference in Long Beach, California. During the session on resilience by design, speakers connected preparedness to the decisions companies make about suppliers, contracts, people and internal communication.

The panel included Loredana Di Salvio, a global trade executive at L’Oréal; David Morrow, who previously worked with the United States Air Force and now runs Supply Chain Engineers LLC; and David Bruder, a senior auditor in contract compliance at Revenew International. Matt Talbert, ASCM’s senior manager of research, moderated the discussion.

Their perspectives pointed to a common challenge: Preparing for disruption requires more than predicting what might happen. It requires understanding what the business depends on, how decisions are made and where the organization can change course.

Know the supply chain before it is disrupted

Di Salvio described five pillars of resilience: supply chain intelligence, communication, processes, people and culture. Her starting point was understanding the network before an event forces the organization to investigate it.

“I think that you don’t have to wait for disruption to know your supply chain,” she said.

She recommended mapping upstream relationships through tier three, understanding partners and manufacturing flows across countries, and quantifying the network’s strengths and vulnerabilities. That work helps establish where action may be needed when conditions change.

But supplier knowledge alone is insufficient. Di Salvio also emphasized the ability to redesign processes, develop flexible employees and build a culture that can respond to changing conditions. People need to understand other functions’ perspectives and manage projects across organizational boundaries, she explained.

Morrow brought the discussion down to individual operating constraints. He recalled a shop where one employee’s testing expertise was essential to output. When that person was absent, capacity fell. The example illustrated a vulnerability that could be overlooked in a risk review focused mainly on external disruptions.

His recommendation was to begin within the area a manager controls and ask what would happen if a necessary resource became unavailable: a person, a material or the power needed to operate.

Make the financial case for preparedness

A recurring theme was the difficulty of explaining resilience investments to executives who see an immediate expense but an uncertain future benefit.

Di Salvio recommended speaking in terms that connect supply chain decisions to leadership priorities: total cost of ownership, cash flow, profit and loss, the balance sheet and customer service. She also urged supply chain leaders to ask colleagues what information they need and what determines success in their functions.

“My suggestion would be always be a good listener, but also ask questions,” she said.

That approach can change how a sourcing decision is evaluated. In discussing tariff-driven supplier changes, Di Salvio cautioned against treating the tariff savings as the entire business case. Qualifying a new manufacturer, establishing transportation routes, negotiating logistics contracts and managing potential quality or delivery problems all belong in the evaluation.

The point was not to avoid changing suppliers. It was to assess the full consequences before responding to pressure for immediate savings, she noted.

Morrow similarly argued that risk should be evaluated against the business activity it could interrupt, rather than simply the purchase price of the affected item. A low-cost component can support a much larger product investment, making the consequences of its absence disproportionate to its unit cost.

He also cautioned that identifying a risk does not automatically dictate mitigation. Companies need a process to identify and assess risks, prioritize which warrant attention, choose a response and monitor the result. Estimates of probability and impact involve uncertainty; they help guide decisions rather than provide a precise forecast.

Treat the playbook as a starting point

The panel did not suggest discarding existing procedures. Instead, speakers emphasized adapting them and learning from the circumstances in which they are used.

A response prepared for one supplier failure may contain useful actions for another, Morrow explained, even when the details differ. The task is to determine what carries over, what needs adjustment and what the plan does not address.

Di Salvio urged organizations to make time for postmortem reviews, despite the pressure to move immediately to the next problem. The review should examine more than the operational result. It should also consider whether employees had the information, tools and working conditions needed to make decisions under stress.

A future disruption may take a different form, she said, but reviewing the response helps develop the flexibility to handle it.

Planning horizons also need to reflect the business. Di Salvio pointed to differences between industries and product lead times, and recommended aligning supply chain plans with manufacturing and sales. Morrow likewise emphasized understanding the pace of the product cycle and watching for constraints deeper in the supplier network.

Use audits to improve the next decision

Bruder added a perspective grounded in looking backward. Contract compliance audits, he explained, can reveal overcharges, improve contract terms and identify process weaknesses that should be addressed before they recur.

“So it’s a historical look back that allows you to plan better in the future,” he said.

His recommendation was to understand where money is being spent and organize that spending into risk areas. Invoice volume and contract structure can help direct attention toward relationships that warrant closer review.

That makes audit findings useful beyond cost recovery. They can inform the information companies retain, the contracts they write and the processes used to check whether suppliers are performing as agreed.

Bruder’s contribution also broadened the discussion beyond a single catastrophic event. Recurring payment or compliance problems can accumulate, requiring attention to current operations alongside preparation for future shocks.

Build readiness without overwhelming the organization

Asked about replacing manual processes with new technology, Morrow advocated incremental change. His advice was to begin with improvements the organization can make using existing resources and prepare employees for the technology that follows.

“Start off small, identify what you can do,” he advised.

The recommendation connected technology adoption to the same organizational capabilities discussed throughout the panel: clear processes, informed employees and the ability to adjust without overwhelming the people responsible for execution.

Di Salvio also encouraged supply chain leaders to communicate opportunities alongside vulnerabilities. Improvements in lead times, inventory and business continuity can help colleagues understand the value of the work without making every discussion about an impending crisis.

Resilience by design, as the panel described it, is an ongoing management practice. It begins with knowing the network and its constraints, connecting decisions to business consequences, and learning from what happens. The playbook remains useful, but the organization’s ability to question and adapt it is what makes it work, they concluded.

Executive takeaways

A June 2, 2025, article titled “The abandonment epidemic is real” on Retail Dive really hit home for me because of something that I have been stating for a long time: in a commoditized world, execution is the new competitive edge.

Cited in the article is a survey by SCAYLE and their interviews with over 1,500 shoppers, which found that 81% of respondents have abandoned a brand in the past year.

The reasons for the abandonment are various and could likely be divided into the main categories of bad policy versus poor performance. These include:

• Different in-store versus online experience
• Different online experience, e.g., computer versus mobile
• Missing discounts
• Confusing prices
• Hidden fees
• Limited payment options
• Confusing, or lack of, return policy
• Inventory that is available … until it is unavailable

At the root of the problem is probably a fragmented system of software applications that may be connected but are not really integrated. Data is likely duplicated without there being a master source of truth. Transactions are transferred from one application to another, and probably between supply chain partners like a 3PL, but the timing is not conducive to supporting the smooth supply chain operational flow. This is all too familiar, as it is typically what I find at brands that are selling B2B/B2C to retailers and grocers as well as D2C. 

We are all consumers of products and services who have undoubtedly abandoned shopping somewhere because of poor execution, whether in-person or online. The world is full of other choices for darn near everything. As you read this article, consider what other choices (a.k.a., competitors) there are for your own company’s products or services. 

If you can deliver the perfect order (see my article series on Supply Chain Management Review here) you will pretty much guarantee that your goods will be available when consumers want them.

How is your company protecting itself against consumer abandonment? Or is this not something that you’ve been thinking about until now?        

Some managers have taken the current geopolitical climate as a signal to stop worrying about sustainability. However, the issues this term encapsulates remain and have, if anything, intensified. Supply chain managers still face climate-related disasters such as droughts or floods that make inputs unavailable. They also need to account for difficulty securing insurance for operations in flood or fire-prone locations, even as customers and other stakeholders continue to sanction firms they perceive as behaving irresponsibly. It may currently be possible to ignore sustainability in some settings, but over the long-term, this is likely to backfire. Sustainability still needs to be addressed.

Coping with relentless disruptions, a changing climate, and a volatile regulatory and political environment means most supply chain managers are already transforming the structure and operations of their supply chains to make them more resilient and environmentally sustainable. It would be natural to put worrying about how the chain impacts and is impacted by people and communities on the long finger. This would be a mistake. Supply chain managers should be putting people at the center of their transition thinking.

Here is the uncomfortable part. We have become extraordinarily good at optimizing supply chains for cost, speed, and flexibility, but that mastery is also the problem. The models that delivered it were built on the assumption that people are a cheap, abundant, and interchangeable resource, to be adjusted as demand dictates. That assumption is now colliding with a shrinking workforce, a more discerning talent pool, tightening regulation, and a wave of new technologies that can either amplify the problem or expose it. The optimized supply chain many managers are proud of is the very thing leaving them most vulnerable.

The roots of this run deep. Globalization, the last radical innovation in SCM, was built on scaling the Toyota Production System into a worldwide model of excellence. This delivered real benefits for firms, but its consequences for people were an afterthought. As the model reached global scale, jobs shifted across the world and the nature of work changed, yet these impacts were barely considered until the model was already entrenched. People, communities, and society were never carefully integrated into the foundations of SCM thinking.

The point we are making is simple: SCM is people-blind. That blindness shows up in the models managers use every day that treat workers as just another resource to be optimized. For example, platforms like Uber create “just-in-place” workers who end up exactly where the service is needed, much as materials arrive just-in-time for the assembly line. But to keep the service cheap for customers and profitable for the platform, that waiting time goes unpaid, shifting the cost onto workers and society.

Remaining people-blind through the current transformation would mean ignoring the underlying demographic trends of an aging and shrinking workforce and the associated war for talent. This would mean repeating the mistakes of the past while missing the opportunities of the present. People are more motivated to work for organizations whose values they believe align with their own and when they sense a disconnect, they may quit or disengage. Tech workers, for instance, are refusing to work on military contracts over the harm they perceive their employers doing. Likewise, issues such as child or forced labor in the supply chain can lead to lost sales and, in some cases, boycotts of goods or services from certain regions or companies.

Given what stakeholders learned from globalization, they will intervene if firms do not proactively address how their current supply chain transformation impacts people and communities. These stakeholders are increasingly concerned with how goods and services are made and where they come from. Workers, communities, and governments are already on alert for how digitization, automation and AI are reshaping supply chain operations and impacting the workforce and the communities where the supply chain operates. Finally, the externalities created by globalization produced legislation such as supply chain due diligence laws, legislation that will increase if the next transformation is deemed harmful.

However, foregrounding people also presents opportunities. The prospect of AI destroying jobs may dominate today’s headlines, but the underlying demographic trends remain. Treating people as part of your supply chain transformation, rather than an afterthought, positions your supply chain for the future. This requires considering people when making supply chain decisions and ensuring the costs of harming or improving social systems visible are to shareholders, supply chain partners, regulators, and wider society. In what follows, we highlight what managers should be considering, and then offer concrete suggestions for building people into their supply chain transformation.

What to consider

It is accepted that a decision in one link of the chain affects the resilience, cost structure, and environmental impact of the entire chain. But supply chain choices about materials, due dates, where work is done and so on also affect people across the chain. When managers outsource to locations with weaker regulatory enforcement, routinely change order volumes or due dates at the last minute, or accept bids that cannot be met within legal limits on pay and working hours, they are implicitly accepting harm to people. Demanding an order to be expedited at a low price is asking the supplier to push workers harder, perhaps through excessive overtime or by cutting corners on health and safety.

These choices determine how attractive your firm is to potential employees, the motivation of the workforce across the supply chain, and the likelihood of regulatory oversight or pressure from customers or other external stakeholders. Ending people-blindness starts with recognizing that most supply chain decisions have a direct or indirect impact on conditions for workers elsewhere in the chain or on the communities in which it operates.

What does supply chain excellence mean now? Transforming supply chains requires changing how we think of supply chain excellence. Supply chain excellence captures the practices (including but not limited to lean) that improve the supply chain. Traditionally, improvement was determined using profit-focused metrics like increased productivity. Going forward, however, sustaining social-ecological systems will be a necessary condition for a supply chain to operate. New models of supply chain excellence will therefore need to treat sustainability as a minimum expectation in the pursuit of profits.

This does not mean that working to make supply chains responsive or efficient should stop, but rather that you need to rethink how to achieve these goals without putting people at risk. Managers must learn to respond to uncertain demand without creating precarious or indecent work. Stakeholders will punish firms whose efficiency and responsiveness requires using people as buffers to create flexibility (e.g., Uber’s just-in-place workers). Consider Zara: some of its early success came from cutting lead-times and increasing flexibility by sourcing close to home, in countries with higher wages and stricter regulations. Rethinking efficiency and responsiveness could start here.

Finally, innovation, especially radical, fast-moving innovation, creates significant uncertainty about how and when work will be done and by whom. That uncertainty increases the likelihood of harming workers. New models of supply chain excellence cannot be developed without simultaneously considering how to integrate people in the innovation process in a way that, at a minimum, does no harm.

This requires a shift in the question we ask. We typically make a business case for protecting people: showing how better conditions reduce turnover or defects, and therefore benefit the firm. What the transition demands is a people case for the supply chain. Not harming the workforce is treated as a necessary condition, in the same way that meeting a quality standard or a delivery window is. The practical difference is enormous. A business case can be overridden the moment the spreadsheet says otherwise; a design constraint cannot.

The hardest part: Scaling up while scaling down. New models of supply chain excellence will need to be scaled. Scaling a new supply chain model is critical for a startup. It is dramatically harder for an established firm, because transitioning to a new model means scaling up a new supply chain while scaling down their current supply chain. For instance, a start-up in the EV space needs to build a new supply chain from scratch, which is no small task. But for an incumbent firm reliant on combustion engine technology, the transition to EVs is harder still. It must build a new supply chain while continuing to operate the old one. That means duplicating many supply chain operations and possibly creating conflicts between the current and new supply chains.

The complications multiply because the profits from the existing chain are usually needed to pay for building and scaling the new one. Top managers tend to frame this as a problem of cannibalizing sales. But the supply chain implications run deeper. A fundamental principle of supply chain management is that unit prices fall as order volumes rise. Equally, once customers know a new technology is coming, they expect a discount to buy the old one. This creates a serious dilemma for a supply chain manger trying to manage a chain out of business while maintaining profits since as volumes of the old technology decline, they need suppliers to cut not raise prices. Worse, they will be doing so when the supplier knows that they won’t be needed at some point in the future. In other words, the supplier also needs to build a new model or find new customers. They will rightfully jump ship as soon as possible. Managing the existing supply chain out of business in an orderly and predictable fashion will be extraordinarily difficult.

Finally, shuttering facilities or winding up an entire supply chain, even one that does a great deal of harm to people, communities or the environment, also means layoffs and disruption. Managers therefore need to consider how to manage a facility or entire chain out of business without leaving lasting scars on the displaced workers and communities where the chain once operated.

Who pays when a relationship ends? SCM thinking treats long-term collaborative relationships as a clear win; the foundation of trust, joint improvement, and innovation. But such relationships are resource-intensive to build and maintain. In a complex supply chain, or one with many suppliers for each input (the many farms behind a food product, the long tail behind a complex assembly), expecting every buyer-supplier relationship to be close and collaborative is unrealistic. And a chain with thousands of small suppliers is not just thousands of relationships; it is hundreds of thousands of workers and the communities they live in, each of whom your decisions reach but none of whom you can see. Managing every relationship well enough to protect the people inside it is not merely expensive, it is structurally impossible. The answer is not better audits. It is to design the chain so that it needs far fewer relationships to manage in the first place; reducing the number of links, shortening the distance to the people affected, and choosing sourcing models that concentrate rather than scatter your relationships. The most powerful thing you can do about unmanageable complexity is to stop building it.

There is another, more uncomfortable question: what happens to those relationships when a supply chain must scale down? A relationship built gradually over decades becomes a liability the moment a supplier’s business model proves unsustainable, or they are unable or unwilling to change with you. The dedicated assets, integrated processes and mutual dependence that made the relationship valuable is what makes it slow and costly to exit. And the exit cannot be managed the way the relationship was. A supplier you are winding down has every incentive to extract what it can while hunting for new customers. Managing a partnership you intend to keep is different from winding one down. But the supplier’s incentives are not the only ones that matter, and this is where people-blind thinking becomes starkest. While the relationship is between buyer and supplier, the cost of dissolving it falls on the supplier’s workers and the community built around them.

Managing an exit in a people-centric way means treating those workers and communities as stakeholders in the exit rather than as collateral damage from it. Workers, their representatives and community stakeholders such as regional development offices where at-risk operations are located, can also offer creative solutions and have exposure to other supply chains going through the same transformation. They want to see your transformation succeed, and they are likely to bring ideas from outside your supply chain. Innovation requires leveraging a network, and workers, their representatives and local communities often have an expansive one containing actors that are not in yours.

What to do: Supply base rationalization

One of the elements of the previous revolution was the rationalization of supply bases. Transforming supply chains means this is going to happen again. A long-term relationship becomes a liability when a supplier is unable or unwilling to rethink how it provides value, and the need to transition means the need to rethink who does what in the chain.

If you have yet to do so, you need to assess which suppliers have the incentives, capabilities, and desire to change with you, and which are unlikely to be part of the chain in the future. And, once more, don’t forget the people. Suppliers who fail to grasp this necessity are likely to limit their exposure to new ideas while increasing the risks their operations pose for the rest of your supply chain.

For suppliers likely to make the transition with you, start the process early. Here you are basically doubling down on the long-term, trusting relationship you have built. This will be (mostly) familiar. Roles will probably change and it may be your suppliers or the supplier’s suppliers who lead the transformation.

Next, some relationships that are currently important or strategic will end in the near-term future. But first, you both likely need to maintain the relationship to help fund your respective transitions. Being open and honest about the direction of travel and leveraging existing trust and communication pathways to find an amicable way to separate may defuse the tensions that come from knowing a relationship is ending.

Finally, the most complex issue will be adding suppliers where a new skill, capability or product is needed. These suppliers, and what they supply, will be unproven. There is no relationship to leverage and no time for the traditional trust-building process. Yet, the need to add a supplier suggests that, at least in the short- to medium-term, this will be a strategic item. Current thinking on strategic items is that they require a single supplier. But some of a new supplier’s unproven innovations will fail, and their willingness and ability to share information and meet performance targets are unknown. Using multiple suppliers will be more effort to manage but will also offer access to a wider range of innovations and suppliers’ networks while reducing the risk of non-conformance.

Taking this a step further, the nature of the relationship is often dictated by the category of spend. To oversimplify, strategic items rely mainly on close, trusting relationships, while leveraged items are typically managed at arm’s length. The need to innovate along with the transitional period of scaling up and scaling down, suggests that even within a category, relationships will vary supplier to supplier. Suppliers who will not be transitioning with you, whose primary incentive is to milk you for profits until they can find new customers, will likely require strong contracts. But relying on the relationship still makes sense for suppliers who have incentives to manage the partnership out of business in a planned and orderly way. A similar calculus applies to the scaling-up process. Truly new suppliers will require more oversight and a larger investment in the relationship. And for those suppliers making the transition with you, the current relationship is one of your most valuable assets, leverage it along with your workers and theirs.

What to do? People in the age of AI

Periods of great uncertainty require picking courses of action without knowing the outcome. People-centric supply chain transformation will require many of these choices. However, there are things you can do now to increase your likelihood of success.

This starts with an increased awareness of how supply chain choices have wider implications for workers and society. Consider those implications before making changes, not after. And while we advocate for building new models of supply chain excellence, the existing models do have attributes worth keeping. For instance, involving the workforce in continuous improvement is foundational to current best practice, because workers know their own jobs best. It would be easy to ignore this during a period of transition, given that the jobs and the skills they require are likely to change dramatically. That would be a mistake. Change efforts need to involve the workforce, or their representatives, wherever possible. AI might allow you to eliminate some entry-level managerial jobs, but the workforce’s accumulated experience is an intangible asset that should not be discarded. And most workers will want to grow and change with you if given the right training and support.

 

Using AI purely for efficiency gains risks further deepening people-blindness. Algorithmic scheduling, automated supplier selection, and demand-driven workforce management make it easy to treat people as a buffer, at a scale and speed no human planner could manage. It is the just-in-place logic taken to extreme.

Yet the same tools can become a practical means of seeing people before a decision is made, rather than after. A digital twin of an operation can model what a last-minute order surge will do to working hours and safety risk at a supplier two tiers down, surfacing a human cost that has always been real but invisible. Building people into the transformation means deliberately writing the human question into the tools themselves. When commissioning a new system or optimizing sourcing, specify worker impacts such as volatility in hours, safety exposure or the load pushed onto a supplier’s workforce as KPIs reported alongside cost and service level, not as an afterthought your managers must search for. A model optimizes what it is told to optimize. If people are not in the objective function, the technology will simply become very efficient at not seeing them, and the people-blindness that was once a limit of human attention becomes a property of the system itself.

Conclusion

For decades the question around sustainability was a defensive one: does it pay to protect the environment, workers and communities? That is the wrong question. The right question for this transformation is sharper and far more uncomfortable: how can we make supply chains efficient, responsive, resilient and sustainable without using people as the shock absorber? Firms that keep asking the old question will keep treating people as a cost to be cut. Firms that ask the new one will design the transition around people from the start.

This is not a plea for softness. It is a clear-eyed assessment of where supply chain risks now lie. The relentless optimization of cost, speed, and flexibility that made the last supply chain revolution a success is what left supply chains exposed. When labor was cheap and abundant, treating workers as an adjustable resource looked like good management. It no longer is. As the workforce shrinks, talent is choosing employers who behave responsibly and customers and other stakeholders are no longer blind to the harm that globalization caused.

The last revolution diffused its model first and discovered its consequences for people only after the model was too widespread to easily change. You have the chance to invert that order, but only if you choose to. For instance, AI can be used to industrialize the old mistakes or to prevent them.

The supply chain managers who will thrive in the coming transition will not be the ones who treat social responsibility as a layer on top of business as usual, they will be the ones who redefine what supply chain excellence means. They will treat the well-being of the people in and around the chain not as an outcome to be measured after the fact, but as a design constraint. Doing so is no longer just the responsible choice. Increasingly, it is the most sound one.

About the authors

Mark Pagell, Ph.D., is the Chair in Global Leadership and a full professor of Sustainable Supply Chain Management at University College in Dublin, Ireland. He can be reached at mark.pagell@ucd.ie.

Miriam Wilhelm, Ph.D., is a full professor of Sustainable Supply Chain Management at Vienna University of Economics and Business (WU), Vienna, Austria. She can be reached at Miriam.Wilhelm@wu.ac.at

Executive takeaways

When a tier-two plant goes down, the control tower lights up. Transit times, inventory positions, and late ASNs appear on a map. Someone opens a war room. The question that decides whether the line keeps running is simpler and usually unanswered: which other part, from which approved source, at which revision, can go into this assembly this week without another engineering change?

That answer does not live in the tower. It lives in the product record.

I work in the digital twin sector after years as a manufacturing and industrial engineer on product lifecycle management, including the path from engineering BOM to manufacturing BOM. The disruptions I have watched stall a factory were rarely a mystery of freight. They were a mystery of master data. The tower could see that a supplier had stopped shipping. It could not see whether the part on the dock was interchangeable, whether the second source was still on the approved list for this revision, or whether last month’s design change had reached the supplier who still held the old print.

Resilience programs keep buying visibility. Visibility without a clean product record is a well-lit view of a shortage you cannot legally or technically substitute.

Three fields the tower cannot invent

Interchangeability is a product attribute, not a buyer’s opinion.

Two items that share a description are not substitutes. Form, fit, function, and the qualification evidence that says they may replace each other belong on the item and on the BOM line with effectivity. When a shortage hits, planning asks manufacturing engineers, who ask design, who opens a folder. Hours become days. The tower’s ETA was never a constraint.

If interchangeability is not a maintained relationship in the product record, every disruption becomes a custom engineering event. That is not agility. That is a queue.

Approved-vendor effectivity is a dated rule, not a spreadsheet tab.

Most plants have an approved vendor list. Fewer have an approved vendor list that knows when a source is valid for which revision of which item. A supplier qualified for revision B is not automatically qualified on revision C after a material change. A second source added for a 2024 program may be expired for the unit you are building in 2026.

During COVID-era and later geopolitical shocks, teams discovered they were placing expedite orders against sources that quality had already pulled. The tower showed a willing supplier. The product record, had anyone queried it as a system of record, would have shown a closed qualification. Resilience that ignores effectivity creates a second problem: a part that arrives and cannot be used.

Change must propagate past the ECO number.

A design change that stays inside PLM is a conversation among engineers. A design change that matters to the supply base is a packet: affected items, new revision, cut-in date or serial, disposition of stock and open POs, and the drawing or model the supplier is allowed to build. If purchasing and the supplier still hold last quarter’s file, you have two products in one part number.

Multi-tier response fails here in a predictable way. Tier one gets the new print. Tier two does not. The component that shows “on time” in the tower is the wrong revision. Incoming inspection becomes the control tower.

What to demand of the product record before the next disruption

Treat the next resilience review as a master-data review. Ask for four proofs, not another dashboard mockup.

• First, can planning retrieve, for a constrained item, the approved alternates and the revisions those alternates may replace without a meeting? If the answer is a shared drive, the record is not ready.
• Second, can procurement see vendor approval as a dated object tied to item and revision, including the date the approval ends? If vendor status is a column that someone updates after the fact, expedites will keep landing on expired sources.
• Third, can a released change list the supply side jobs it created—drawing to supplier, PO retarget, stock disposition—and show which of those jobs are closed? Approval in PLM is not propagation.
• Fourth, can the manufacturing BOM and the planning item agree on what will be kitted if the alternate is used? An alternate that exists only on the engineering bill will not appear in the MRP run that is supposed to save the week.

These are not IT niceties. They are the difference between a substitution that ships and a substitution that becomes a quality escape.

Put the tower on top of the record, not in place of it

Control towers earn their keep on transit, capacity, and inventory movement. They do not know whether two screws are the same screw. That knowledge is designed, qualified, and dated in the product record. When the record is thin, the tower’s recommendations look decisive and fail at the dock.

A practical sequence is unfashionable and short. Pick the fifty items that stop the line. For each one, clean interchangeability, vendor effectivity, and the last two changes that should have reached the supply base. Measure time from “supplier down” to “approved alternate on an open order.” That time is your real resilience KPI. The map on the wall can wait until that number moves.

Supply-chain resilience is often sold as seeing the disruption sooner. Seeing it sooner is useful. Being allowed to respond- with a part that is interchangeable, a source that is still approved, and a change that arrived at the plant that makes it- is the work. That work starts with the product record.

About the author

Akarshan Gulhane works in the digital twin sector at PTC. He previously worked at the intersection of manufacturing and product lifecycle management as a manufacturing and industrial engineer, including PLM implementation. His work focuses on the handoff from eBOM to mBOM to executable work instructions, and on why digital models stall before they reach the cell.

Today is Friday, Oct. 16

1. IEA accelerates diesel release as supply pressure persists

IEA members agreed to release more than 20 million barrels of diesel and accelerate remaining March commitments as disrupted energy flows keep transport-fuel prices near record levels.

Reuters  |  Read the original article

2. First copper enters U.S. strategic minerals stockpile

Mercuria delivered 12,000 metric tons of Congolese copper to Project Vault, marking the first reported metal added to the U.S. government’s planned strategic critical-minerals reserve.

Financial Times  |   Read the original article

3. U.S.-Mexico tariff negotiations lose momentum

Negotiations over U.S. tariffs on Mexican automobiles, steel, aluminum and derivative products have reportedly stalled, prolonging major uncertainty for manufacturers planning cross-border sourcing and production.

Reuters  |   Read the original article

4. Chinese buyers expand U.S. soybean bookings

Chinese state buyers booked another 10 to 13 Panamax cargoes of U.S. soybeans, bringing the week’s reported purchases to roughly 1.2 million to 1.5 million metric tons.

Reuters  |  Read the original article

5. ASCM puts geopolitical volatility atop its 2027 outlook

ASCM’s latest trends outlook ranks geopolitical volatility ahead of agentic AI while emphasizing that governance, cybersecurity, talent, traceability, automation and regional sourcing must be managed as connected priorities.

Logistics Management  |   Read the original article

6. Electronics supply chains return to longer planning horizons

Memory scarcity, component constraints and changing trade rules are pushing electronics manufacturers toward multiyear planning, closer supplier collaboration, longer contracts, alternative materials and broader sourcing networks.

Supply Chain Dive  |   Read the original article

7. Prologis raises warehouse-development outlook

Prologis increased its 2026 development-starts forecast after third-quarter results beat expectations, citing healthy demand from e-commerce, advanced manufacturing and the expanding global data-center supply chain.

FreightWaves  |    Read the original article

8. Unilever ties digital-twin expansion to measurable outcomes

Unilever plans to expand AI-enabled digital twins from five factories to about 40 sites, using defined operating goals, repeatable deployment methods and workforce training to guide adoption.

Supply Chain Dive  |   Read the original article

9. Sustainability commitments rise while governance trails

MIT’s 2026 survey of 1,810 supply chain professionals found record support for sustainability, but tariff disruptions, limited cross-functional governance and weak board oversight continue to impede execution.

MIT News  |   Read the original article

10. Local AI targets maritime contract risk without the cloud

An MIT capstone proof of concept runs locally to compare maritime contracts with approved templates, identify possible demurrage exposure and link findings to source language for expert verification.

Supply Chain Management Review  |   Read the original article

Executive takeaways

Editor's Note: The SCM thesis From Unstructured PDF Files to Contract Intelligence: A Local LLM RAG Pipeline for Extracting Demurrage Terms and Conditions, was authored by Stephen Day III and Wathila Ekanayake and supervised by Dr. Elena Dugundji (elenna_d@mit.edu). For more information on this research, please contact the thesis supervisor.

Our project tackled a critical bottleneck in maritime shipping: the inability to rapidly evaluate and negotiate incoming shipping contracts. We focused our proof of concept on the largest cost driver in these agreements: demurrage. In the oil, gas, and chemicals industry, profit lives in the margin between commodity prices and operational efficiency. Few costs erode that margin faster than demurrage, which is the penalty charged when loading or unloading a vessel runs past the allowed time. The terms governing these penalties, including laytime rules and exceptions, are buried in dense maritime charter agreements that routinely run 40 to 60 pages in length.

Our sponsor, a large energy company, signs thousands of these shipping agreements annually. Reconciling a single proposed voyage to understand demurrage exposure often takes an analyst up to six hours of reading redlined legal language and cross-referencing conflicting terms. Our tool revolutionizes this deal desk workflow. It uses semantic AI to let analysts instantly audit incoming, unsigned agreements against baseline templates. Because these maritime contracts are highly confidential, we had one hard constraint: the entire system had to run locally.

The cost of slow evaluations

A single voyage is governed by a hierarchy of documents. The Charter Party is the foundational template of standardized terms. The Recap is the authoritative record of what was negotiated and overrides any conflicting language elsewhere. Riders and addendums layer in further amendments.

When it takes hours to manually reconcile this complex legal hierarchy, discrepancies are usually caught only after the vessel has sailed. This delay completely neutralizes any leverage to negotiate or enforce a contractual amendment.

The core technical problem, then, was not simply using local AI to read PDFs. It was instantly converting an incoming, unstructured agreement into a structured data asset that faithfully respects this legal hierarchy. This speed allows traders to catch risks before the ink dries, all in an environment where confidentiality is non-negotiable.

A data center in a lunchbox

The instinct in enterprise AI is to assume that serious generative reasoning demands expensive third-party cloud infrastructure or data-center GPUs. Our project demonstrates that this assumption is overstated, and we are not the only ones making the point. At CES 2026, AMD CEO Lisa Su held up a lunchbox-sized mini PC and ran a 235-billion-parameter model on it live. This demo did not rely on a data center, the cloud, or a rented GPU. That machine was built on the same AMD Ryzen AI Max+ 395 “Strix Halo” processor that powers our system.

The contract intelligence system runs on a single compact Minisforum MS-S1 Max workstation built around that Strix Halo processor with 128 GB of unified memory. By permanently dedicating 96 GB of that memory to AI workloads at the firmware level, our machine holds a 26-billion-parameter language model and an 8-billion-parameter embedding model in fast local memory at the same time. The result is performance comparable to an enterprise server, sitting quietly in a standard office, requiring no climate-controlled server room and no external network connection.

That last point is the strategic one. For a firm handling sensitive commodity contracts, keeping the entire intelligence-gathering process on-premises is not a nice-to-have; it is what makes the technology viable at all. The economics reinforce the case. Because the entire AI stack is open source, there is no vendor lock-in and no per-token billing. Once the hardware is purchased, the marginal cost of analyzing one more contract is simply the electricity to run it.

From raw PDFs to a structured asset

Turning a contract into intelligence happens in a single on-device ingestion pass. Text is extracted with precise character coordinates, and, critically, strikethrough redlines left over from negotiation are detected and stripped out. Deleted clauses that remain visually present in a final PDF would otherwise be embedded into the knowledge base and silently bias the tool toward obsolete terms. Detecting where vector lines cross the midline of a character, and dropping that text, proved to be a mandatory step rather than a refinement.

The cleaned text is then semantically chunked, tagged with metadata that encodes document type and hierarchy level, and indexed for both keyword and semantic search. In parallel, the system runs a “map-reduce” pass in which the language model reads the full agreement and distills every clause, obligation, liability cap, and financial threshold into a structured “clause inventory” of typically 80 to 250 lines per contract. This information is stored in a local database. Decoupling this structured inventory from raw text retrieval was one of our biggest breakthroughs: it lets the system audit an entire contract directly and completely, rather than hoping a keyword search surfaces the right needle in a 60-page haystack.

Three tools, one secure dashboard

Users reach everything through a single web interface that hides the underlying infrastructure. Behind it sit three tools:

• Fact Retrieval answers natural-language questions about contract terms, combining keyword and semantic search to return precise results to questions about a given contract. Each factual claim the AI returns is based on the source text, and a clickable in-text citation is generated. It resolves hierarchy automatically: when a Recap and Charter Party conflict, the Recap wins.
• Contract Compare audits a proposed agreement against an approved baseline. This produces a structured report of missing protections, weakened obligations, newly added risks, altered financial thresholds, and an overall risk verdict.
• Executive Summarize condenses an entire agreement into a standardized seven-section brief covering parties, freight rates, laytime and demurrage, key obligations, liability caps, notable amendments, and a risk assessment, so leadership can triage the business logic of a deal without reading 50 pages of legal language.

Trust through verification

Accurate insight is useless if an analyst cannot confirm it. Commercial traders and legal experts, reasonably, will not stake a counterparty dispute on an answer they cannot trace to a source. So every claim the system makes carries a citation, and every citation is clickable. Selecting one opens a locally hosted document viewer, and the user’s attention is called directly to the exact paragraph the answer came from. Bridging the gap physically, back to the original text in its original context, was the decisive factor in moving the system from an experiment to a tool people trusted.

From reactive to proactive

The payoff is a shift in posture. A review that once took an analyst up to six hours now runs in moments, which is fast enough to apply to every proposed agreement as it arrives, not just the ones someone has time to scrutinize. Instead of discovering an unfavorable term after a vessel has sailed, traders can audit each incoming contract against an approved template before signing. They enter negotiations armed with evidence rather than institutional memory, keeping unfavorable clauses out of the active fleet entirely. The value compounds with every contract reviewed: faster, more consistent commercial scrutiny that protects voyage economics … all without a single byte of data ever leaving company-controlled hardware.

Today is Thursday, Oct. 15

1. SK Hynix links U.S. NAND investment to China equipment access

SK Hynix is seeking U.S. approval to equip unused NAND capacity in China while connecting that production plan to a proposed Solidigm memory-chip factory in the United States.

Reuters  |   Read the original article

2. TSMC says AI demand still exceeds chip capacity

TSMC said AI-related demand continues to outstrip available capacity as it raised its 2026 revenue-growth outlook and evaluates additional U.S. manufacturing expansion beyond its Arizona complex.

Reuters  |   Read the original article

3. WTO expects supply chain diversification to endure

The WTO’s director-general said repeated wars, tariffs and chokepoint disruptions are producing costlier alternative trade routes that some companies will retain as a hedge against future shocks.

Reuters  |   Read the original article

4. Digital twins move from visualization to decision testing

Supply chain digital twins can help teams simulate disruptions and compare operational responses, but their value depends on reliable data, validated models, focused use cases and clear AI boundaries.

Supply Chain Management Review  |   Read the original article

5. UK extends defense logistics modernization contract

The United Kingdom extended Leidos’ defense logistics contract through June 2029, continuing support for more than 70,000 commodity stock numbers and distribution to over 600 locations worldwide.

Leidos  |   Read the original article

6. Missouri logistics expansion adds 830,000 square feet

Wagner Logistics plans two leased facilities totaling more than 830,000 square feet in Liberty, Missouri, supporting a regional manufacturer’s warehouse and production-support program and creating 79 jobs.

City of Liberty, Missouri  |    Read the original article

7. Procurement AI advances through reversible decisions

TransAlta supply chain leader Liza Tsimbal describes using AI for contract review, drafting and scenario planning while favoring reversible, supervised decisions before granting systems greater autonomy.

Zycus  |   Read the original article

8. ABB expands precision robots and vision-based teaching

ABB introduced two precision-manufacturing robots and an AI-assisted training capability that it says can shorten commissioning and cycle times by reducing conventional programming requirements for manufacturers.

Robotics 247  |  Read the original article

Executive takeaways

For most of the last decade, a supply chain “digital twin” meant a pretty picture: a 3D model of a warehouse or a network map with dots that lit up when something went wrong. Useful for a steering committee slide, less useful for a planner trying to decide whether to reroute a container at 11 p.m.

That version of the twin is being retired. What is replacing it is closer to a live, queryable model of the entire network—nodes, flows, constraints, and now, increasingly, an AI layer that can reason over it—built not to be looked at, but to be asked questions and acted on.

That shift matters because it changes what a digital twin is for. A static model tells you what your network looked like when someone last updated it. A live twin tells you what will happen three moves from now if a key port closes tomorrow—and, in the newest deployments, it can propose or even execute the response.

From a picture of the network to a model you can interrogate

The distinction worth holding onto is between visualization and simulation. Visualization answers “where is everything right now?” Simulation answers “what happens if?” A control tower dashboard is visualization—valuable, but it still requires a human to translate what they’re seeing into a decision. A true digital twin is a model with enough fidelity in its constraints—capacity, lead times, contractual minimums, transportation lanes—that it can run a scenario and return a defensible answer: reroute through this port and you add four days and $80,000; hold at the current port and you risk missing the promotional launch window.

Analysts tracking the space this year describe digital twins moving from an emerging capability to what Deloitte has called strategic decision infrastructure—the layer leaders use to test constraints and understand ripple effects before committing resources, rather than a monitoring tool bolted on after the fact.

Why this is arriving now

Three things converged to make live twins practical rather than aspirational. The data foundation caught up first: cloud ERP, integrated planning suites, and IoT-connected assets now feed a twin continuously instead of through a nightly batch job, so the model reflects the network as it actually is, not as it was yesterday. Second, generative AI made twins easier to build and query—a planner can now describe a scenario in plain language rather than needing a simulation specialist to code it. Third, and this is the part getting the most attention from vendors this year, agentic AI is turning the twin from something a person interrogates into something that can act: diagnosing a developing bottleneck and triggering a response inside pre-approved boundaries, without waiting for someone to open the model first.

None of this required a single breakthrough. It required the pieces—clean data, cheaper compute, better reasoning models—to mature at roughly the same time, which is why so many organizations are evaluating twins simultaneously rather than one industry leading and others following years later.

Where the early value is landing

The clearest wins so far are narrow and operational, not enterprise-wide. A manufacturer uses a part-level twin to catch a design change that would have quietly increased scrap rate. A distribution network runs nightly what-if simulations against a demand spike and pre-positions inventory before the spike hits, rather than reacting to a stockout report. A logistics team models a port closure against every active lane in minutes instead of the days a manual re-planning exercise used to take. In each case, the twin isn’t replacing the planning process—it’s compressing the time between a signal appearing and a defensible option being on the table.

That compression is the real ROI case, and it’s a more honest one than the efficiency-percentage claims that circulated a few years ago. A twin that shaves two days off a re-planning cycle during a disruption is worth more, in a bad month, than one that trims half a percent off routine transportation cost in a good one.

The adoption gap: pilots are easy, scaling is not

Despite the momentum, most organizations are still further along in piloting twins than in running them as production infrastructure. The obstacles are familiar to anyone who has tried to stand up an enterprise data initiative: fragmented systems that don’t agree on a single version of inventory or lead-time truth, a shortage of people who can build and maintain simulation models, and an unresolved question of who owns the twin once it stops being a project and starts being a system multiple teams depend on.

The organizations moving fastest are treating the twin less like a one-time IT deployment and more like a product with a roadmap—starting with one high-value use case (a single facility, a single lane, a single category), proving the model’s fidelity against real outcomes, and expanding scope only once the underlying data and governance can support it. Trying to model the entire global network on day one is the most common way these initiatives stall.

What supply chain leaders should ask before investing

A digital twin is only as useful as the decision it’s built to support, which argues for starting from the decision rather than the technology. Where is the cost of a slow or wrong decision highest—a single-source category, a capacity-constrained facility, a lane with frequent disruption? Does the underlying data have enough fidelity to model that decision credibly, or would the twin be simulating against guesses? And critically, once the twin can recommend or trigger an action, who is accountable for that outcome, and what’s the boundary of what it’s allowed to decide without a human in the loop—the same governance question agentic AI is forcing across every other corner of supply chain execution.

The forecasting era gave supply chains better eyes. Digital twins, done well, are giving them a rehearsal space—the ability to run the disruption before it happens instead of only reacting once it does. That is a different kind of value than a more accurate number, and for organizations still treating the twin as a dashboard rather than a decision engine, it’s the gap worth closing next.

About the authors

Souraj Roy is an MBA student in Supply Chain Management at Symbiosis Institute of Operations Management (SIOM), Nashik, with prior experience at Infosys. He holds SAP Certified Associate—SAP IBP for Supply Chain and Kinaxis Certified Maestro Author Level 1 certifications.

Samriddha Basu is an MBA student in Operations & Supply Chain Management and interned at CEAT. He holds the Certified Associate in Project Management (CAPM), SAP Certified Associate—SAP IBP for Supply Chain, and is a Kinaxis Certified Maestro Author Level 1.

Today is Wednesday, Oct. 14

1. Renewed attacks deepen Hormuz shipping risk

At least five incidents involving commercial vessels were reported around Hormuz in 48 hours, extending severe security, insurance and routing risks for energy shipments through the critical waterway.

Reuters  |  Read the original article

2. Aluminum prices conceal persistent physical shortages

Gulf smelter outages remain well below prewar output while low exchange inventories and elevated regional premiums signal physical tightness, despite increased Chinese and Indonesian exports filling part of the gap.

Reuters  |  Read the original article

3. AI demand pushes ASML toward further capacity expansion

ASML raised its annual sales outlook and is evaluating additional production increases as semiconductor customers signal demand for lithography equipment extending into 2029 and 2030.

The Wall Street Journal  |   Read the original article

4. Hyundai commits $15.5 billion to India manufacturing

Hyundai Motor Group plans to invest $15.5 billion in India through 2032, expanding Hyundai and Kia production capacity while adding research, engineering and supply chain activity.

Reuters  |   Read the original article

5. Canal constraints redirect cargo toward Los Angeles

Suez and Panama Canal concerns are shifting an estimated 5% more cargo toward Los Angeles, where September volume reached 1.04 million TEUs amid drought risk and higher fuel costs.

Supply Chain Dive  |   Read the original article

6. Uber Freight searches for scale and profitability

Uber Freight is showing volume and multi-service customer growth under CEO Rebecca Tinucci, but the logistics unit remains unprofitable as it expands internationally and applies AI to transportation management.

The Wall Street Journal  |  Read the original article

7. AI-assisted prototyping moves closer to operations

Generative AI can help operations managers convert local knowledge into working prototypes for overlooked planning and execution problems, allowing teams to test usefulness before seeking larger enterprise investments.

Supply Chain Management Review  |  Read the original article

8. Deep-learning vision targets variable robotic picking

OSARO says SightWorks combines pretrained models, RGB and 3D data, and end-effector awareness to help robotic arms handle unfamiliar products without extensive SKU onboarding or manual programming.

Robotics 247  |   Read the original article

Executive takeaways

The fall has a way of filling the calendar before you have a chance to think about what belongs on it. There are immediate operating demands, year-end targets and decisions about what to put in next year’s budget. And before you know it, another year is upon us.

Taking time away requires a good reason.

That is why I believe next week’s NextGen Supply Chain Conference deserves a place on your calendar. And you get can get 25% off the regular registration fee if you use the code NGReader25 at checkout.

The conference opens Wednesday, Oct. 21, at the W Nashville and runs through Oct. 23. As you consider your 2027 priorities, it offers an opportunity to hear how other organizations are approaching decisions you may soon need to make, and to ask questions before those decisions become commitments.

Which technology investments deserve to move forward? What needs to change before an initiative can scale? Where will your team need new skills? Those questions have consequences for budgets, service and the people responsible for delivering results.

Put your investment assumptions to the test

A promising technology demonstration is one thing. Understanding what it will take to make that technology useful in your operation is another.

Our agenda creates room to examine both. Wayfair’s Nitin Kapoor will join me for a fireside conversation about the company’s logistics evolution and the future of home delivery. Eli Lilly’s Dr. Mar Gimeno will explore how agentic AI challenges established supply chain leadership habits.

Other sessions get into the details that can determine whether an investment succeeds. Nordstrom will examine the data-quality problems uncovered by questioning an aged-inventory metric. Polaris and Cycle Labs will discuss the role of testing in protecting supply chain systems.

Together, those perspectives can help attendees ask better questions about their own plans: Are we addressing the right problem? Can we trust the information behind the decision? What must work reliably before we add another layer of technology?

Connect transformation to operating decisions

Next year’s priorities will need to hold up in daily operations.

Southern Glazer’s Wine & Spirits and Dematic will share their fulfillment transformation, connecting automation with capacity and network design. GE HealthCare’s Sam Colen will discuss a finance-first approach to testing and scaling AI for inventory decisions. Amazon’s Debanshu Sharma will present work on predictive carrier-risk scoring.

These are opportunities to explore how organizations connect an idea to an operating need. The value for attendees lies in understanding the choices behind an approach, the conditions it requires and the questions to consider before applying it elsewhere.

Our small group discussions provide additional opportunities to examine those details with speakers and fellow attendees. A focused conversation may help you refine a project, spot an assumption that needs testing or identify someone whose experience can inform your next step.

Make people part of the plan

Technology plans also create workforce decisions.

Friday’s talent panel brings together perspectives from Johnson & Johnson and Schneider Electric, with University of Tennessee’s Dan Pellathy moderating, to examine education, training and career pathways.

If your 2027 plans include new systems, greater automation or different responsibilities, the capabilities of your team belong in that discussion from the beginning.

That is the purpose behind our theme: Innovate. Upskill. Transform. Progress requires attention to the tools, the people and the operating changes that connect them.

There is still time to join us

I would like you to leave Nashville with something you can use: a sharper question for your team, an idea worth investigating or a connection that helps you move a decision forward.

Thank you to the sponsors and speakers who make these conversations possible. Now I hope you will be part of them.

NextGen opens next Wednesday, Oct. 21. Take a look at the agenda, identify the discussions that matter to your organization, and register to join us at the W Nashville, Oct. 21–23. Use the coupon code NGReader25 to get 25% off your registration fee.

Before you finalize next year’s priorities, give yourself an opportunity to sharpen them.

Executive takeaways

Editor’s Note: The views and opinions expressed in this article are solely those of the author and do not reflect the views of Amazon or any of its affiliates.

The next operating advantage may come from tools no enterprise software team would ever choose to build. Not because those tools lack value. They may solve a meaningful planning problem, reveal an overlooked constraint, or eliminate hours of recurring work. But they are often too local, too specialized, too uncertain, or too small to compete for scarce product and engineering capacity.

Operations and supply chain organizations have lived with this constraint for years. Managers closest to the work recognize opportunities that existing systems do not address. They know which data should be combined, which exceptions matter, and which decision would improve if the right information appeared at the right moment. Yet many of those ideas never become products. They remain spreadsheets, manual routines, or requests waiting for a development roadmap.

Generative AI is changing that equation.

For the first time, an operations manager can move directly from understanding a problem to creating the first working version of its solution.

That is more consequential than faster software development. It changes who has the power to create operational capability.

From software customer to product creator

Historically, the operations manager was a customer of technology. The manager described a need. A product manager translated it into requirements. Analysts examined the data. Engineers developed the application. Designers shaped the experience. The process was rational, but expensive. It required organizational attention before anyone knew whether the idea would work.

AI compresses many of those capabilities into a development partner that can be directed through natural language.

A manager can explain the operating problem, challenge the proposed design, test different analytical approaches, create an interface, and refine the product as understanding improves. Cloud infrastructure, accessible databases, and AI coding tools provide the remaining foundation.

The manager does not need to become a software engineer. The manager needs enough applied AI fluency to recognize what can be built and enough operational judgment to direct what should be built.

This creates what I call the AI-Augmented Operations Leader: an operations or supply chain manager who can convert operational knowledge into working digital tools without waiting for the traditional product-development cycle.

Making operational knowledge executable

The most valuable operating knowledge is rarely captured fully in a process map or requirements document.

It lives in experience.

A supply chain manager may know that late inbound appointments are not the real problem. Congestion appears only when several conditions coincide: a particular shipment mix, limited dock flexibility, longer-than-normal unloading times, and labor commitments elsewhere in the operation.

Existing systems may track each variable independently without answering the decision that matters: Where is congestion likely to emerge, and which action will prevent it?

Previously, the manager might maintain a manual model or request a planning application that never became a sufficient enterprise priority.

Now the manager can direct AI to combine the available data, test the operating logic, and create a custom application around that decision. The first version may be imperfect. That is precisely why rapid prototyping matters. The manager can use it, learn from it, and improve it while the problem is still current.

The breakthrough is not the dashboard or prediction itself.

It is that the manager’s understanding of the operation can become executable.

Evidence before investment

Traditional product development asks an idea to prove its importance before it exists.

The proposal must attract sponsorship, earn funding, enter a roadmap, and survive competition from larger initiatives. Small or uncertain opportunities are filtered out early, even when they may create meaningful operational value.

AI reverses this sequence.

Managers can build before they persuade. They can test before they seek investment. They can demonstrate usefulness through actual operating experience rather than projected benefits.

Evidence can precede organizational conviction. 

This opens a previously underserved space between Excel and enterprise software. It is a space for purpose-built operational products: more capable than spreadsheets, more adaptable than standardized systems, and inexpensive enough to justify experimentation.

Some will be temporary. Some will solve a problem at one facility and disappear. Others will expose a need shared across the network and grow into larger products.

Their initial value comes from no longer requiring certainty that they deserve to exist.

A new source of operating advantage

The most important effect of AI in operations may not come from one enterprise platform or one predictive model. It may come from thousands of managers gaining the freedom to build around the realities they see every day.

They can create metrics that existing systems never anticipated. They can turn recurring analyses into applications. They can test alternative operating mechanisms. They can build around local constraints without waiting for those constraints to become enterprise priorities.

Product and engineering teams will remain essential for large-scale systems. But they no longer need to be the starting point for every operational idea.

That distinction matters.

When software creation was scarce, the organization’s technical capacity determined which operational ideas could be explored. As that scarcity recedes, the constraint shifts elsewhere.

The advantage will belong to managers who can identify which parts of their operational knowledge are worth turning into products—and to organizations that recognize this new form of leadership.

For years, operations managers were expected to work within the capabilities of the systems available to them.

AI is giving them the power to shape those capabilities themselves.

Today is Tuesday, Oct. 13

1. Tanker scarcity strands heavy crude in the Americas

Surging tanker costs and nearly full U.S. refinery utilization are trapping heavy crude in North and South America while refiners in Asia and Europe struggle to secure diesel-rich feedstocks.

Reuters  |   Read the original article

2. Medical supplier links U.S. expansion to tariff protection

BD plans $3 billion in strategic U.S. production investments and five billion additional annual units of medical consumables under an agreement offering conditional relief from future Section 232 tariffs.

Supply Chain Dive  |   Read the original article

3. Large industrial projects test U.S. manufacturing ambitions

Proposed steel and shipbuilding investments in Iowa and Philadelphia face questions about labor, space, customer demand and economics, illustrating the execution risks between industrial announcements and sustainable production.

The Wall Street Journal  |   Read the original article

4. ASCM puts geopolitical volatility atop its 2027 agenda

ASCM’s 2027 trends outlook ranks geopolitical fragmentation and trade-policy volatility first, followed by agentic AI, while linking technology progress to talent, data governance, cybersecurity and operating resilience.

Supply Chain Management Review  |   Read the original article

5. Global manufacturers accelerate materials purchasing

GEP’s September index found stronger raw-material and component buying across major industrial economies, with North American capacity pressures rising and average global logistics costs increasing for the first time since April.

GEP  |   Read the original article

6. Levi’s directs tariff refunds toward logistics and distribution

Levi Strauss plans to allocate roughly one-third of $60 million in tariff-refund reinvestment to distribution and logistics, with most of the broader spending scheduled for its current quarter.

Supply Chain Dive  |   Read the original article

7. Walmart opens automated West Coast fulfillment hub

Walmart opened a 900,000-square-foot Stockton, California, fulfillment center using automated storage and retrieval to compress its order process from 12 steps to five and expand regional shipping capacity.

Supply Chain Dive  |  Read the original article

8. Humanoid order books outpace proof of productive deployment

A Robotics 247 analysis distinguishes announced orders and factory capacity from installed fleets, urging buyers to evaluate accepted units, productive hours, uptime, human intervention and post-sale integration support.

Robotics 247  |   Read the original article

Artificial intelligence remains a major force in supply chain transformation, but it did not take the top trends spot in ASCM’s outlook for 2027. Geopolitical fragmentation, tariffs and trade-policy volatility led the discussion, reflecting a shift toward building operations that can function through continuing uncertainty.

The 2027 Top 10 Supply Chain Trends were presented at the recent ASCM CHAINge conference in Long Beach, California, by Matt Talbert, ASCM’s senior manager of research, and Armand van Oostrom, the trends subcommittee chair and a senior lecturer at The Hague University of Applied Sciences.

The full list:

1. Geopolitical fragmentation, tariffs and trade-policy volatility
2. Agentic AI and autonomous supply chain orchestration
3. Workforce transformation/AI literacy and the talent gap
4. Data foundations & AI governance
5. Cybersecurity & Nth-party digital risk
6. End-to-end visibility & multi-tier traceability
7. Physical AI, robotics & warehouse automation
8. Enterprise/integrated risk management & anti-fragility
9. Supply chain as a revenue & value engine
10. Regionalization, multi-sourcing & distributed production networks.

The ranking offered more than a collection of separate priorities. Van Oostrom emphasized that the trends influence one another and that companies need to consider how they fit together within their own operations.

“What is my top priority and how does it impact all the other ones?” he told the packed room.

For executives, that question changes how the list can be used. A technology project may depend on better data, different employee skills and new security practices. Choosing one priority still requires attention to the capabilities that support it.

Build volatility into the operating model

Van Oostrom described the leading trend as a change in how companies approach uncertainty. The issue is moving beyond reacting to volatility and incorporating it into the way the supply chain is designed.

“You cannot foresee the future, so you better build it in a way that is resilient and agile,” he said.

That perspective connects the first trend with regionalization and multiple sourcing, which ranked 10th. The presenters discussed how disruption and dependence on a single source can drive companies to reconsider the location and structure of their supplier networks.

Van Oostrom added that regionalization can also be part of a network optimization exercise in its own right. Companies may have several reasons to make a sourcing change, and those reasons need to be assessed in the context of the operation.

The risk-management discussion carried a similar message. Talbert argued that a playbook cannot simply remain on a shelf until the next interruption. Companies need to watch immediate risks, scan farther ahead and update contingency plans as conditions change.

“Playbook needs to be dynamic,” he said.

Van Oostrom linked that adaptability to employee preparation and decision-making speed. A company can maintain a playbook while also training people to respond as events unfold. The point is to make the response usable before a lengthy internal process delays action.

Separate AI ambition from operating readiness

Talbert said the subcommittee had deliberately moved beyond treating AI as a single broad category. Breaking it into more specific themes was intended to make the outlook more useful to practitioners deciding what to do next.

The agentic AI discussion focused on the movement from recommending actions toward executing them. Talbert referenced separate ASCM research conducted with the Digital Supply Chain Institute and said automatic execution remained rare, while insights and recommendations were more common.

Autonomous orchestration is a direction of development, rather than a description of how most supply chains already operate, they noted.

The presenters connected progress in AI with data quality and governance. Talbert described a company taking a measured approach through small use cases, testing them and expanding as it learned. Van Oostrom similarly favored controlled experiments that could inform larger deployments.

Workforce readiness was another dependency. The speakers discussed how jobs and processes need to evolve alongside the technology, and how education and ongoing learning can prepare people for those changes.

They generally described AI as an augmenting tool, but the question-and-answer discussion added an important point. Talbert said respondents to one study anticipated the elimination of some customer-service and lower-level inventory roles, creating a potential challenge for entry-level workers.

Van Oostrom emphasized critical thinking and process knowledge when discussing how people should learn to use AI. Training needs to help employees understand the problem they are solving and evaluate the output, alongside learning how to operate the tool.

“More positively, people are still the ones that make the technology and processes work,” he noted.

Connect visibility with security and action

Cybersecurity appeared in the discussion as an operating risk that crosses organizational boundaries. Van Oostrom described the dependence companies have on the security of connected partners and urged collaboration with suppliers and customers.

Talbert called for stronger connections between supply chain and IT teams, noting that an IT interruption can become a supply chain interruption. The concern places security within the service-continuity discussion, alongside physical disruptions to production or transportation.

End-to-end visibility was also linked to what companies can do with the information they obtain. Talbert described an evolution from control towers used primarily as monitoring dashboards toward centers that support action.

The intended change allows professionals to spend less time chasing shipments and more time managing exceptions. Van Oostrom connected visibility with the practical requirements of handling a recall, when the ability to identify affected products and locations becomes especially important.

The robotics and automation discussion extended the same process focus. Van Oostrom said companies should consider the capabilities technology makes possible when designing work, rather than only looking for an existing task to automate.

That approach ties automation back to workforce transformation. A redesigned process changes how people participate in the operation, which makes training and change management part of the investment.

Show the value that supply chain creates

Another theme concerned how supply chain leaders demonstrate their contribution to the business. The presenters argued for a broader view of supply chain as a source of value, including service improvements and opportunities to support customers and markets.

Van Oostrom urged leaders to maintain relationships with disciplines such as finance and engineering. Talbert recommended documenting interventions that prevent disruptions or preserve service, so that their contribution becomes visible to senior management.

His examples were illustrative: Finding an alternative supplier or averting a disruption may save substantial time and money, yet remain largely invisible outside the team that handled the problem.

The recommendation was to track those contributions and use them to make the case for supply chain participation in executive decisions. It connects resilience with the value discussion by asking leaders to explain what their capabilities protect and enable.

Use the trends to set an agenda

Talbert described a four- to five-month development process involving more than 35 subcommittee participants. Members gather material, extract possible trends and help narrow the results before a final vote. The group aims to look three to five years ahead, although Talbert acknowledged that immediate concerns can influence the outcome.

The list is developed new each year, he said. Changes in wording or ranking therefore need to be interpreted with care: A broad theme can split into several more specific issues, and an older priority can reappear in another form.

That point surfaced in an audience question about sustainability, which did not make the 2027 top 10. Talbert acknowledged its continuing importance. Its absence from the ranking should not be read as evidence that companies no longer face sustainability concerns, he said.

The presenters said a fuller report, including more industry detail, was planned for December.

Executive takeaways 

Increasingly, businesses and consumers prioritize delivery speed and reliability when deciding where to buy. That’s why improving the customer experience is logistics leaders’ top priority through 2030, according to a recent Gartner survey.

Brant Seaton

This strategic focus is reinforced by a separate Gartner study, which found that 90% of logistics leaders now report to the C-suite, including 30% who report directly to the CEO.

Two retail companies highlighted in this article exemplify this shift. Both use integrated logistics services from a single provider to reduce costs, speed freight movement, and scale inventory and operations with seasonal demand changes.

Their success in managing warehousing and fulfillment through a single, coordinated logistics network helps them move freight faster, more efficiently, and with greater flexibility.

Customer expectations are driving supply chain strategy

The Amazon Effect has made fast, transparent, and accurate delivery the standard for consumers and businesses alike.

One of the companies is Outdoor Cap, a Bentonville, Arkansas-based headwear manufacturer. Its customers are asking the same three questions, says Peel Chronister, head of warehouse and logistics: How easy is it to do business with you? How quickly can you deliver? Will my order arrive complete and on time?

The second company is a high-volume seasonal retailer facing the same expectations under a different business model. Every year, more than 5 million cartons move through its kitting network phase from June through August before shipping to stores nationwide.

Both businesses are meeting customer expectations with more than faster transportation. They use an integrated logistics network that can scale with demand, execute consistently, and keep products moving without interruption.

Execution creates competitive advantage

Execution is a defining competitive advantage in supply chain management.

For the high-volume seasonal retailer, execution involves receiving imports, processing inventory, coordinating transportation, and replenishing stores during a compressed selling season, where every day saved creates additional selling opportunities.

Few retailers experience demand spikes this dramatic, but manufacturers, distributors, and retailers alike face constant pressure to respond quickly to promotions, product launches, seasonal demand, and shifting customer behavior.

Creating a competitive advantage in fast-changing environments requires eliminating delays across the supply chain.

Outdoor Cap shows that its competitive advantage in fulfillment speed and cost doesn’t come from adding warehouse space alone. It took redesigning its logistics network to eliminate unnecessary handoffs, reduce delays, and keep products moving from supplier to customer.

Design logistics networks around flow

How freight flows through supply chains is giving companies a competitive advantage. Outdoor Cap worked with a new logistics partner to redesign its freight flows, reduce costs, and improve responsiveness.

Import containers arriving at the Ports of Los Angeles and Long Beach are immediately cross-docked by its logistics partner and transloaded into intermodal containers and domestic trailers, minimizing unnecessary handling and accelerating freight movement.

One key advantage is that approximately 2% to 3% of its customer orders, including custom products and time-sensitive West Coast shipments, are “trapped” or fulfilled directly from California by the same logistics partner rather than routing all freight to Arkansas.

Today, Outdoor Cap’s integrated logistics network receives about 400 import containers annually. In 2026, that volume is expected to increase to between 450 and 500 containers, as the company’s scalable logistics model supports continued growth.

Technology strengthens the network

Technology is essential for managing a supply chain, but it alone is not the source of competitive advantage.

Artificial intelligence, warehouse management systems, transportation management systems, and real-time visibility platforms deliver value only when supported by a well-coordinated logistics network.

Outdoor Cap’s technology strategy reflects that. By integrating its ERP directly with its logistics provider’s warehouse management system, the company has real-time visibility across inventory, fulfillment, and transportation in a single operating environment.

That visibility enables direct fulfillment from California and both rail intermodal and truckload transportation to Arkansas.

Technology strengthens the integrated logistics network by improving coordination and execution. Without the infrastructure and underlying support of the logistics provider’s network, the technology would merely provide visibility.

Competitive advantage starts with an integrated network

Both companies mentioned here rely on integrated warehousing and fulfillment managed by a single logistics provider to move freight faster, reduce costs, and scale as customer demand changes.

Their supply chain strategies reflect the broader shift by manufacturers, retailers, and distributors to build lasting competitive advantage through integrated logistics networks.

For organizations looking to strengthen their supply chains, the first step isn't choosing a warehouse, implementing new software, or adding transportation capacity. It's designing a network that can execute consistently at scale for moving products faster and more efficiently.

About the author

Brant Seaton is President of Knight-Swift Supply Chain, which provides integrated warehousing, distribution, fulfillment, and transportation services across North America, backed by the scale and resources of Knight-Swift Transportation.

Today is Monday, Oct. 12

1. Hormuz traffic falls as military action reaches commercial shipping

Strait of Hormuz vessel traffic fell to a two-month low after tanker attacks intensified and the U.S. military disabled a Panama-flagged cargo ship it said attempted to breach the Iran blockade.

Reuters  |   Read the original article

2. Russian diesel deal offers limited relief from fuel pressure

The United States temporarily eased sanctions as Russia agreed to supply more than 300,000 metric tons of diesel immediately, although analysts questioned whether the volumes could materially reduce near-record fuel prices.

Reuters  |   Read the original article

3. Los Angeles port handles a record September surge

The Port of Los Angeles processed 1.04 million TEUs in September as holiday imports, Panama Canal restrictions and high fuel costs redirected cargo toward the shorter trans-Pacific route.

Reuters  |  Read the original article

4. EU and China outline an initial trade compromise

European and Chinese trade officials reached a preliminary agreement that could lower selected tariffs, stabilize rare-earth supplies and limit Chinese hybrid-vehicle exports, but implementation still requires approval from EU leaders.

Associated Press  |  Read the original article

5. Walmart’s warehouse automation push meets operational limits

Walmart’s multibillion-dollar warehouse automation program has encountered reliability, packaging and cost challenges as the retailer works to automate complex item handling across a distribution network serving 5,200 U.S. stores.

The Wall Street Journal  |  Read the original article

6. U.S. robotics growth still depends on Chinese suppliers

American robotics developers remain heavily dependent on China’s component ecosystem, manufacturing scale and engineering speed, complicating efforts to build domestic capacity while competing for leadership in physical AI.

The Washington Post  |  Read the original article

7. Scaling predictive AI requires operating-model changes

An SCMR contributor argues that predictive-AI pilots often stall because data, decision ownership and performance incentives remain unresolved, recommending that companies scale proven use cases incrementally rather than transforming entire networks at once.

Supply Chain Management Review  |  Read the original article

8. Industrial parts test exposes AI accuracy gaps

A vendor-designed test found leading general-purpose AI models answered no more than 52% of 100 industrial-parts questions correctly, highlighting the need for verified product data and human review.

Robotics 247  |   Read the original article

Executive takeaways

Supply chain leaders have been running predictive AI pilots for approximately three years: demand sensing, dynamic safety stock, disruption detection, control tower simulations among others. Most of them have been successful in that they met the success criteria set by the organizations for each pilot.

However, an additional insight from these projects is that many of those same initiatives never graduated beyond a pilot project. Thus, the model that could predict potential supply chain risks still sits on a dashboard. The demand-sensing initiative that was able to increase forecast accuracy more than 10 points has yet to be rolled out to the remaining twelve.

The “follow up” to the news that “AI will come and disrupt the way you do business with your suppliers,” is that AI has already done so, in many ways. The question is no longer about “can we get this tech to work?” but rather “why have so few of those companies who were willing to spend a lot of money on this capability actually captured any value from it?” Those factors create such significant barriers to success for many of these firms when compared to what seems like an easier path to successful implementation by other firms; those are likely to be among the hottest topics of discussion regarding the future of supply chain management in 2026.

The scale-up gap is real, and it’s not about model accuracy

Most vendors of predictive intelligence solutions claim that their models have proven effective within pilot environments, but executives who initiated the pilots often tell a different story and it’s not about the model. The true challenges to the success of these models are as follows: making sure that a workable model is incorporated into operational planning, determining which party has ownership of the decision when there is disagreement between the algorithm and the planner, and evolving from the single product line or region level proof of concept to support all products and regions in an enterprise network.

Over the last couple of years, industry studies have shown that most AI pilots in the supply chain/operations area are reporting positive results; however, only a small percentage of these pilots will be successfully deployed (within the next 12 months) as an enterprise-wide solution. The problem lies in the gap between proof of concept and full-scale enterprise solutions.

Figure 1: The fork in the road. Most predictive AI pilots clear their success criteria — the divide between stalling and scaling happens after that point, and is decided by organizational factors rather than model performance.

Root cause one: data readiness was never really solved, just worked around

Pilots typically tolerate some level of unclean data. A single business group with a single curated dataset is common in pilots as is a data science team manually cleaning inputs over the course of several months. The results of these efforts are frequently very compelling. However, it is difficult to replicate this type of performance when scaling. Feeding the same predictive models data coming from a dozen different ERP systems, varying SKU hierarchies by region and supplier records that have never had to be normalized due to nearly a decade of M&A can produce much lower levels of performance.

The businesses that can scale in a successful manner do so by treating data infrastructure as an output from the pilot, rather than the model. They identify where they have failed with respect to the foundational nature of their data through the pilot, and then fund the unappealing (but necessary) work of creating clean master data and establishing integration pipelines before announcing success with the pilot, and then scaling.

Root cause two: nobody owns the decision the algorithm makes

The forecasting model provides a simple way to understand how much safety stock should be added for a certain part by telling you to add two weeks of safety stock. The recommendation may be easy to test in a pilot, but putting it into production requires clear accountability. There has to be someone inside your organization accountable when the prediction ultimately turns out to be incorrect. Many organizations experience what I call an ownership gap, which happens with their AI system. Procurement thinks planning owns the problem, planning thinks finance is involved (due to potential working capital impacts), and finance was never at the table during the model development.

This is not a data science issue and therefore doesn’t have a data science solution; this is about decision rights and governance that determines which recommendations may be executed automatically, which require human approval, and who owns the resulting decisions. Decision rights and escalation rules should be established during the pilot, before recommendations enter production.

Root cause three: the KPIs still measure the old world

Most supply chain organizations measure planners, procurement teams, and logistics managers on reactive and operational KPIs such as on-time delivery, cost per unit, and inventory turns. Predictive AI is about taking action in time to prevent a disruption scenario, which is a fundamentally different scorecard than what most organizations use today, and one that is much harder to quantify.

If we do not have new incentives for our businesses, they will not be able to generate the changes in behavior necessary to achieve the benefits of predictive AI. There is just as much rework involved with creating the right metrics as there is in writing code to implement successful scaling of AI within a supply chain organization.

What actually works: an expansion model, not a transformation program

The companies that have made the leap from piloting to enterprise predictive applications didn’t get there by undertaking a broad-based, sweeping change initiative. Rather, it appears that the companies which have made the most progress in deploying predictive applications have taken the following steps:

• Take it slow and start small by focusing on a limited set of business problems for which there is demonstrable value, rather than jumping right in with a broad-based initiative that promises to revolutionize the company’s use of information technology.
• Budget for the data and integration issues as part of the initial pilot project, and view the resolution of these problems as the real project—the construction of the predictive models is merely a means of exposing these issues.
• Designate decision governance rules and ownership prior to implementation, rather than waiting until there is a crisis. Explicitly define when automated decisions will be made and when human decisions will be needed, and make sure that someone is accountable for the outcomes of these decisions.
• Revise the compensation and performance evaluation system to reward and recognize individuals who take action on the predictive models. Turn the models’ outputs into actions.
• Expand a proven use case to additional sites, product lines or regions before adding substantially different applications.

Figure 2: The scaling playbook. Most companies that scale from pilot to enterprise-level predictive AI appear to follow this progression, prioritizing governance, data, and incentives as integral components of the implementation process.

Conclusion: the next competitive divide isn’t who has the best model

The organizations that stand to benefit from predictive AI applications likely will not be those with the most advanced algorithms. What differentiates one company from another is the ability to create the governance, data infrastructure, and incentive structures that allow predictive analytics to drive behavioral change effectively.

The move toward predictive supply chains from reactive ones, debated extensively throughout the industry last year, was always the correct direction to head. However, it remains to be seen which organizations will be successful in executing the change. The next 18 months will see a bifurcation between the companies that have approached AI as a pilot project and those that have embraced it as an operational transformation, with repercussions for both sides.

About the author

Kishan Kumar is a buyer at Pearse Bertram based in Bloomfield, CT. In his nine-year career, he has worked on strategic sourcing, procurement and supply chain management within the manufacturing, industrial automation and FMCG industry. Mr. Kumar earned his MBA from Southern Connecticut State University as well as his Supply Chain Strategy & Management certification from the Indian Institute of Management. Mr. Kumar is a certified Six Sigma Green Belt, SAP Certified Implementation Consultant, Agile Leader Certified professional. Kumar is currently an active professional member of the Association for Supply Chain Management (ASCM) and has contributed to SCMR.com, Techtarget & ExpressPharma. He can be reached at kumarkishan95@gmail.com or via LinkedIn.

Today is Thursday, Oct. 8

1. Tanker costs weaken U.S. crude’s appeal in Asia

Record tanker rates are making U.S. crude uneconomic for some Asian refiners, encouraging buyers to consider Middle Eastern and Latin American alternatives as disrupted routes and longer voyages reshape energy sourcing.

Reuters  |   Read the original article

2. Logistics index reaches a four-year high

The September Logistics Managers’ Index rose 3.6 points to 70.2, signaling accelerating activity as tighter warehousing and transportation capacity pushed logistics costs higher across the market.

Logistics Management  |   Read the original article

3. Extended import peak lifts port outlook before slowdown

An extended import peak is lifting 2026 container volumes at major U.S. ports, but the latest Port Tracker outlook anticipates softer inbound demand as the year closes.

Logistics Management  |   Read the original article

4. 7-Eleven weighs bringing supply chain work in-house

7-Eleven’s North American business is considering bringing more sourcing and distribution work in-house as inflation, fuel expenses and other supply chain costs pressure margins and consumer spending.

Reuters  |  Read the original article

5. AI demand lifts WTO trade forecast

The WTO raised its 2026 merchandise-trade growth forecast to 3.9%, crediting demand for semiconductors and AI infrastructure, while cutting its services outlook as higher fuel costs constrain transportation and travel.

Reuters  |   Read the original article

6. CBP begins processing selected tariff refunds

CBP has begun processing refunds for certain finally liquidated entries affected by invalidated IEEPA tariffs, although current eligibility depends on whether importers supplied valid identification by the July deadline.

Supply Chain Dive  |   Read the original article

7. Amazon adds two China-to-U.S. airfreight options

Amazon introduced two China-to-U.S. airfreight services offering either regional inventory placement in seven to 10 days or a lower-cost option with estimated transit times of 11 to 15 days.

Supply Chain Dive  |   Read the original article

8. Factory backlogs grow as input pressures mount

Unfilled U.S. factory orders increased 0.6% to $1.61 trillion in August while shipments were nearly flat, as manufacturers reported mounting pressure from steel, diesel, freight and tariff costs.

FreightWaves  |   Read the original article

9. Global container volume sets an August record

Global container traffic reached a record 17.46 million TEUs in August, up 3.7% from a year earlier, while a commercial pricing index climbed to its highest level since July 2024.

FreightWaves  |  Read the original article

Executive takeaways

Global supply chains are increasingly exposed to shocks and disturbances, which are becoming more costly. Businesses are operating in an environment where disruptions can emerge simultaneously and compound one another. Tariffs, geopolitical tensions, weather events, labor constraints, transportation capacity, cybersecurity threats, and demand volatility can all affect the same network.

A Business Continuity Institute 2024 report found that almost 80% of organizations’ supply chains were disrupted over the past 12 months, with most experiencing between one and ten disruptions. According to DHL Supply Chain’s Insight 2030 research, 70% of supply chain executives expect cybersecurity threats to disrupt their operations by 2030. Nearly as many are bracing for rising labor costs (69%), labor shortages (66%), natural disasters (63%), and geopolitical turbulence (62%).

Given how this underscores the need for resilient supply chains, it’s not surprising that a growing number of companies are actively stress testing their networks. The exercise critically enhances and strengthens traditional supply chain planning. A good stress test evaluates how the supply chain would perform under a range of disruptive, high-impact scenarios to determine the operational, financial and customer experience consequences. The goal of the exercise is to expose vulnerabilities and single points of failure before an actual disruption forces a company to discover them in real-time.

Resilience stress testing provides companies with a forward-looking, data-driven view of how crises would impact operations, and enable proactive resilience improvements rather than reactive fixes. For instance, it can reveal where inventory buffers are insufficient, where warehouse capacity becomes constrained and where seemingly minor disruptions could create significant downstream consequences.

How companies can stress-test their supply chain

Stress testing doesn’t have to begin with a costly, time-consuming, enterprise-wide implementation. Companies can start by identifying portions of the supply chain where disruption may have the greatest operational impact or where there are suspected vulnerabilities. Once the scope is determined, the exercise typically starts with a baseline digital representation of the supply chain that covers suppliers, capacities, inventory levels, lead times, and logistics flows.

Defined stress scenarios are then applied to this model, and any performance changes are noted and measured. Unlike general scenario planning, which is often qualitative, resilience stress tests use quantitative simulations to produce quantitative outputs on such things as expected stockouts, time-to-recover, and cost impact, which leaders and AI systems can use to make important decisions.

Simulation scenarios should include both probable (high-likelihood, low-impact) and extreme (low-likelihood, high-impact) events to effectively balance day-to-day cost optimization with long-term business survival. This can include a wide range of disruption events, including the following shocks that reflect real-world events that have repeatedly disrupted global supply chains in recent years:

• Loss of a key supplier or manufacturing site
• Multi‑week port closures or customs delays
• Regional conflicts or political unrest
• Labor shortages
• Sudden shifts in customer demand
• Cyberattack taking down a supplier’s system
• Increase in freight rates
• Weather event like hurricanes, tornados or winter storms

The most effective exercises include involvement from multiple functions and roles. Procurement may understand supplier dependencies, while logistics knows transportation alternatives. Operations understands production constraints, while finance can quantify margin and working-capital impacts. Sales and customer service understand which customers and orders may be most sensitive to delays, while HR can best navigate labor issues. And, of course, IT has vital knowledge and insight when it comes to cyberattack vulnerabilities and impacts.

Technology continues to make stress testing more powerful

Advanced technology has evolved supply chain stress testing from static planning to dynamic, real-time resilience. This ongoing evolution is creating an active, protective layer that automatically flags vulnerabilities and enables resources to be shifted more quickly. 

Digital Twins

Digital twin technology has emerged as the leading method for stress testing complex supply chains by creating virtual replicas of physical networks. It helps companies identify vulnerabilities and evaluate mitigation strategies before problems arise in the actual supply chain.

By simulating different scenarios in the digital twin network, companies can carry out resilience stress testing to see how logistics operations respond to a range of disruption scenarios. The technology enables supply chain and logistics planners to easily modify and rerun scenarios to test various strategic adjustments. These simulations provide a comprehensive view of potential impacts on service levels, the cost to serve and supply chain integrity without affecting real-world operations.

Digital twins are quickly moving from single-use applications to utilization in more comprehensive systems, integrating all aspects of logistics operations. Larger 3PLs are leveraging digital twins to enable data-driven decisions that enhance efficiency, resilience and sustainability across the operation. It provides the ability to clearly define tradeoffs and easily understand the implications on total cost to service. In many cases, they can use the technology to quickly model and move to execution in near real time. As the technology matures, its impact is expected to grow as organizations build more interconnected digital ecosystems.

Advanced AI and analytics

Advanced AI and analytics add a further dynamic layer to resilience stress testing, helping ease deployment and usage of the digital models. AI-powered analytics can ingest and process large volumes of unstructured logistics-related data (e.g., geopolitical shifts, weather patterns, supplier financial health) to identify vulnerability patterns and forecast potential disruptions.

The integration of machine learning allows planners to run thousands of complex shock scenarios simultaneously without disrupting live operations. Advanced algorithms can simulate how a hyper-specific event, such as a port closure or a critical component shortage, cascades through multiple tiers of suppliers. By mapping any and all interdependencies within the supply chain, they expose single points of failure that traditional manual audits often miss. This provides a much more accurate picture and offers strong support to demand forecasting and inventory optimization. They can also help identify alternative solutions with natural language processing (NLP).

Because these AI-powered systems continuously learn from each simulation and real-world disruption, logistics providers are increasingly applying them to operational planning. This is further helping to create supply chains that are more diversified, flexible and nimble to better withstand future crises and minimize disruptions.

Supply chain control towers

Supply chain control towers (SCCTs) can connect resilience stress testing directly to day-to-day operations. Instead of running a stress test once a year and putting the results on a shelf, companies can establish continuous monitoring and trigger predefined scenarios when conditions change.

Similar to a physical control tower, an SCCT is a digital platform that offers end-to-end visibility and control over the supply chain. Through software that analyzes data on a business’ upstream and downstream activity, an SCCT acts as a centralized viewpoint that streamlines the way the movement of goods is managed.

These digital hubs integrate data from various sources, including suppliers, logistics providers, and internal systems, to create a single view of operations that provides real-time insights and facilitates proactive decision-making. Enhancing coordination impacts every level of the supply chain—from procurement and production to distribution and customer service. With SCCTs, businesses have minimized disruptions and optimized performance.

When utilized effectively, SCCTs can be a game-changer when it comes to creating a more resilient supply chain. By identifying trends, inefficiencies, and opportunities for improvement, the platform equips businesses with the ability to make informed decisions and refine their supply chain strategies to ensure competitiveness.

A valuable resource during stress testing

Conducting a comprehensive supply chain resilience stress test can often be daunting for companies due to deep data silos, limited modeling tools, internal blind spots, and the resources and technologies needed to adequately conduct the test. Additionally, most companies are still barely scratching the surface when it comes to the supply chain operational value that can be achieved through gathering data and establishing data analytics processes. This is all where a fourth-party logistics (4PL) provider can be a valuable resource.

A 4PL is a strategic partner that manages, operates and optimizes an entire end-to-end supply chain, using service providers and suppliers on a neutral basis and enabling transparency and continuous improvements to provide a company’s logistics needs. A 4PL does not necessarily replace a 3PL relationship. Rather, it can integrate services of different 3PLs and transportation service providers for multiple operations with the goal of optimizing logistics cost.

Because 4PLs manage the entire supply chain and operate across multiple providers, carriers, modes, facilities and geographies, they can provide an insightful and unique view of transportation and fulfillment risks and threats. A 4PL can help weigh best practices, model alternative transportation strategies, identify capacity constraints, evaluate warehouse options, and develop contingency playbooks that can be activated when predefined triggers occur.

With supply chains becoming more complex, disruptions are going to happen. It’s vital that companies understand what will happen when they occur and how they can respond quickly enough to minimize that disruption. Resilience stress testing helps companies realize the concept of “being more resilient” and translates it into a set of measurable performance standards. Advanced technology also makes the exercise more powerful and accurate. By quantifying the trade-offs between costs and resilience, organizations can make informed strategic and operational decisions to better prepare for real-world events.

About the author

Mark Kunar is CEO of DHL Supply Chain North America, overseeing operations across the United States and Canada and a workforce of 52,000 associates. Since joining DHL in 1996, he has held leadership positions spanning operations, finance, strategy and commercial management. Before becoming CEO, Kunar served as chief financial officer and chief strategy officer for the North American business.

Today is Thursday, Oct. 8

1. Germany blocks Cosco deal over supply chain security

Germany stopped Cosco’s planned purchase of an 80% stake in logistics provider Zippel, citing national-security concerns and the risk of greater foreign dependence across critical European supply chains.

Reuters  |   Read the original article

2. Maersk raises inland fuel surcharge in Britain and Ireland

Maersk will increase its emergency fuel surcharge for inland transportation in the United Kingdom and Ireland to 20% on Oct. 12 as conflict-driven energy costs intensify logistics pressure.

Reuters  |  Read the original article

3. North American trade review advances to public comment

The U.S. Trade Representative opened the next phase of the USMCA review, seeking public input after the United States declined to extend the agreement during this summer’s joint review.

Supply Chain Dive  |   Read the original article

4. Zim lifts outlook as container demand strengthens

Zim raised its 2026 earnings forecast by 72%, attributing the revision to stronger demand and higher freight rates as shipping markets continue responding to capacity constraints and geopolitical disruption.

FreightWaves  |  Read the original article

5. Air cargo buyers shift toward shorter contracts

Three-month agreements represented 60% of new airfreight contracts beginning in the third quarter, up from 47% in the second, as shippers sought flexibility amid volatile rates and capacity.

Supply Chain Dive  |   Read the original article

6. Intermodal growth outpaces commodity rail gains

U.S. intermodal traffic increased more than three times as quickly as commodity carloads in the latest week, while metals, petroleum and forest products supported continued gains across freight rail networks.

FreightWaves  |   Read the original article

7. Freight and inputs push McCormick’s inflation outlook higher

McCormick raised its fiscal 2026 inflation forecast to as much as 7%, citing freight, logistics, packaging and other input costs while using pricing, productivity and sourcing actions to offset pressure.

Supply Chain Dive  |   Read the original article

8. Autonomous wildfire system begins California field testing

FireDome opened U.S. operations at Yuba County Airport, where it is testing an autonomous platform designed to detect emerging fires, target threats and deploy retardant capsules before commercial rollout.

Robotics 247  |   Read the original article

Executive takeaways

When we build the NextGen Supply Chain Conference agenda, we start with the challenges supply chain leaders are trying to solve. How do you move beyond an AI pilot? Where does automation improve the operation? How do you prepare people for a different kind of work? And how do you turn a promising idea into a business result?

Bringing the people who can answer those questions together takes a commitment from many organizations. Our speakers share their experiences. Attendees invest their time. Our Advisory Board helps shape the program. And our sponsors provide the support that makes the gathering possible.

[Click to register for NextGen]

As we prepare to meet Oct. 21–23, 2026, at the W Nashville in Nashville, Tennessee, I want to recognize those sponsors and thank them for investing in the NextGen community.

As someone who helps build this program, I know how much support it takes to turn an agenda into a gathering where people can learn from one another. Their support reaches across the conference, from recognizing supply chain achievement to sharing implementation lessons and creating time for conversations over breakfast, lunch and breaks. Those contributions help us deliver a program built around this year’s theme: Innovate. Upskill. Transform.

Recognizing the organizations moving supply chain forward

A special thank you goes to Zion Solutions Group, our Diamond sponsor and sponsor of the NextGen Supply Chain Awards programs on both Thursday and Friday.

The awards recognize organizations advancing intelligent transformation, autonomous operations and collaboration. Thursday’s End User awards include Mars Snacking (Intelligent Transformation), Southern Glazer’s Wine & Spirits (Autonomous Operations), and Ryder and BJC HealthCare (Partnership in Execution). Friday’s Solution Provider awards recognize Netstock (Intelligent Transformation), Pickle Robot (Autonomous Operations) and Robust.AI (Startup). Tractor Supply will also be recognized through our Visionary Award keynote.

Zion's support helps us create a setting in which attendees can learn from the work behind those achievements. Drew Eubank, Zion’s executive vice president and co-founder, will moderate Thursday’s End User awards program.

I am also grateful to our Platinum sponsors, Gather AI and SoftBank Robotics. Their support is an important contribution to the overall conference experience.

SoftBank Robotics is sponsoring Wednesday’s welcome reception in the W Nashville's Living Room lounge, where attendees can meet fellow supply chain leaders, begin making connections and enjoy a performance by Nashville singer-songwriter Nick DeLeo.

Gather AI is sponsoring Thursday’s networking lunch at Zaytinya, where attendees will have time to continue the morning’s conversations and enjoy a performance by Nashville singer-songwriter Emma White.

Case studies that bring the implementation into the room

Our six Gold sponsors—Cycle Labs, Dematic, Dexory, Geekplus, The Modern Data Company and Zimark—are contributing to one of the most useful parts of NextGen: Thursday’s small-group case studies.

[Click to register for NextGen]

These sessions pair technology providers with practitioners to explore the operating challenge, the implementation and the lessons learned. The small-group format gives attendees room to ask questions and examine how an approach might apply to their own organization. Sessions repeat in the afternoon, providing another opportunity to attend.

The Gold-sponsored case studies include:

• Cycle Labs and Polaris: Where test automation and deterministic validation remain essential as companies introduce AI into supply chain systems.

• Dematic and Southern Glazer’s Wine & Spirits: Modernizing beverage distribution through an automated fulfillment operation designed to support growth.

• Dexory and ODW Logistics: Moving from periodic manual warehouse audits toward continuous inventory intelligence.

• Geekplus and Neovia Logistics: Applying shelf-to-person robotics to improve fulfillment performance and exploring the practical work of scaling automation.

• The Modern Data Company and Weatherford: Building shared, governed business definitions that allow AI agents to provide answers traceable to their sources.

• Zimark and Vitti Logistics: Using computer vision to improve inventory and shipping accuracy for high-value goods.

Taken together, these sessions address a question that runs through the conference: What does it take to make technology useful in the operation? I appreciate these sponsors and their practitioner partners for bringing that discussion into the room.

Supporting the conversations between sessions

Thank you as well to Purolator, our Silver sponsor; Corvus Robotics and Verity, our Bronze sponsors; and Argano and AutoScheduler, our Associate sponsors.

Verity is sponsoring Thursday’s breakfast, helping attendees start a full day of learning and discussion. Corvus Robotics is sponsoring Thursday’s morning break, creating another opportunity to connect with speakers, colleagues and potential partners.

Every sponsor’s contribution matters. A conversation during a break can help clarify something heard on stage. Lunch can introduce two people working through the same challenge. Those exchanges are part of what attendees take home from NextGen.

A program worth coming together for

Alongside the sponsored case studies, the program features keynotes, fireside chats, presentations and panels across logistics and fulfillment, retail, food and beverage, and chemicals and pharmaceuticals.

I will sit down with Nitin Kapoor, Wayfair’s VP/GM of Technology, Data and Innovation, to discuss the evolution of its logistics network and the recently launched Wayfair Delivery Plus. Tractor Supply’s Tracie Haislip will accept the Visionary Award and join me to discuss how the company uses its supply chain to support growth. Dr. Mar Gimeno, associate vice president of U.S. supply chain at Eli Lilly, will address the new mindset required for agentic AI in the Friday morning keynote.

[Click to register for NextGen]

Attendees will also hear perspectives from companies including Fanatics, Penske Logistics, Amazon, GE Healthcare, DP World, Target, GXO Logistics, Apple, DHL Supply Chain, Nordstrom and Stanford Medicine. Topics range from forecasting and warehouse execution to data quality, workforce development and the decisions that connect technology with business performance.

And we will make time to enjoy Nashville. Nick DeLeo will perform at Wednesday’s welcome reception. Emma White will perform during Thursday’s networking lunch. Scooter Carusoe, the songwriting name of Travis Hill, will perform at Thursday evening’s reception at the W Nashville’s Butterfly rooftop venue.

Join the NextGen community

To every sponsor: Thank you for helping us bring this program and this community together. Your support makes it possible to create space for learning, recognition and the conversations that carry ideas beyond the conference.

If you have not yet registered, I hope you will join us in Nashville. Review the latest agenda, choose the discussions most relevant to your business and come ready to share the challenges you are working on.

To view the latest agenda, click here

To register for the conference, click here

Organizations interested in sponsoring the conference, click here

Executive takeaways

Global supply chains have entered a period in which uncertainty is no longer a temporary deviation from normality. Armed conflicts, trade restrictions, changes in tariff policy, disruption of established routes, sanctions, shortages of transport capacity, port disruptions and tensions between the world’s largest economies can no longer be treated as isolated emergencies. For international business, they are gradually becoming part of the operating environment.

Catherine Sharapova

Companies are responding by investing in digitalization, forecasting, risk management systems, safety stocks and supplier diversification. All of this is necessary. Yet discussions about supply chain resilience often overlook one factor that cannot be replaced by software, an additional warehouse or a carefully prepared contingency plan.

That factor is the human being.

More precisely, it is the professional who, when the established model stops working, does not wait for the old conditions to return but starts building a new one.

Waiting is also a decision

Consider a typical crisis situation. An established transport route becomes unavailable. The carrier cannot confirm lead times. The supplier is waiting for instructions. The customer is demanding a new delivery date. Finance wants to understand the potential losses first. Legal teams are reviewing the restrictions. Management is waiting for more information.

At this point, two approaches almost always emerge inside an organisation.

The first is: “We need to wait until the situation becomes clearer.”

The second is: “What options do we have right now?”

The line between these two approaches is, in many ways, the line between apparent and real supply chain resilience.

From a human perspective, the instinct to stop in the face of danger is entirely understandable. Our biological response to threat still follows the familiar pattern: fight, flight or freeze. In a corporate environment, the third response often sounds far more respectable than it does in nature. It is called “waiting for additional information,” “avoiding a premature decision,” “obtaining further approval,” or “seeing how the situation develops.”

The problem is that the supply chain does not freeze together with the meeting.

The vessel continues moving. Space on an alternative sailing disappears. Container equipment is repositioned. Production consumes the remaining stock. The customer moves closer to a shutdown point. Another buyer reserves the supplier’s available capacity. A freight rate available in the morning may no longer exist by the evening.

In procurement and logistics, therefore, waiting is not the absence of a decision. It is a very specific decision with a measurable cost.

In some business functions, postponing an issue until tomorrow may change very little. In supply chains, one day can mean the difference between a normal delivery and a production stoppage, between an acceptable freight rate and emergency transportation at any cost, between a stocked shelf and an empty one.

Time quite literally becomes money.

The professionals who take ‘Impossible’ apart

In this environment, people with a very particular professional habit become especially valuable: they do not accept the word impossible as the final outcome of an analysis.

• The route is closed. Which other ports can be used? Can the transshipment point be changed? Is another mode of transport available? Can the shipment be split? Which items are genuinely critical and need to move faster than the rest?
• The carrier refuses to accept the cargo. Who else handles this category? Can the packaging be changed? Can the transport leg be redesigned? Can one route be divided into several segments?
• The supplier says the order cannot be produced on time. What exactly is constraining the lead time: raw materials, equipment, production capacity, components, packaging or testing? Can part of the order be completed first? Can individual items be moved to another production site? Can the production sequence be changed? Can the most critical items be shipped earlier?

For a strong professional, the first impossible does not mark the end of the problem. It signals the need to ask a better next question.

The second and third impossible rarely stop them either. They keep breaking the problem down until the constraint becomes specific and measurable. Once a constraint is specific, it can often be bypassed, compensated for, redesigned around or consciously accepted.

There is, however, an important difference between a merely active person and a genuinely strong professional.

The latter does not seize the first solution simply because it is technically possible.

The first option may be too expensive. The second may be too slow. The third may create an unacceptable quality risk. The fourth may simply move the problem further down the chain.

So the search continues.

A strong specialist develops several options, compares lead times, cost, risk and practical executability, combines solutions and only then selects the final approach. The objective is not to find the first possible answer. It is to find the first answer that meets their professional standards.

These are the people who, during a crisis, create routes and supply models that did not exist the day before.

Why the battle-tested often matter more

There is an old saying: one battle-tested person may be worth two who have never seen battle. In global supply chain management, periods of instability make that idea almost literal.

The ability to act in a crisis cannot be learned entirely from a textbook or a procedure. A substantial part of it develops through repeated exposure to environments in which uncertainty is normal.

A professional who has spent years working within a stable system may be excellent at process optimization, cost reduction, forecasting and inventory management. These are highly valuable capabilities.

But a stable environment rarely teaches someone how to rebuild a broken supply chain in a matter of hours.

A very different professional reflex develops in places where economic storms are not exceptions but part of everyday operations. Where a route available yesterday is closed today. Where a carrier changes its conditions. Where a supplier misses the production schedule. Where currency restrictions change. Where an established documentation process suddenly stops working. Where an alternative has to be created before anyone officially acknowledges that there is a problem.

This kind of experience develops one particularly important quality: the absence of psychological dependence on a single correct scenario.

For such professionals, a supply chain is not a rigid structure. It is a system that can be dismantled and rebuilt.

This is also why people who have been through several genuine crises often see options that would not occur to others. They already know that the sentence “we have never done it this way” says almost nothing about whether it can be done tomorrow.

Resilience is built both before and during a crisis

Of course, no organization should build its supply chain around individual heroics. If the same problem requires a heroic intervention every month, that is no longer resilience; it is poor management.

But it would be an equal mistake to assume that well-designed procedures can fully replace professional judgement.

A contingency plan is useful for as long as the crisis resembles the scenario that was anticipated. The real test begins when two or three assumptions behind that plan fail at the same time.

That is when someone has to create a solution from incomplete information.

The best professionals do more than put out the fire. After the crisis, they turn the workaround into a new capability for the company: they retain the alternative supplier, formalize the additional route, reconsider safety-stock levels, introduce new control points, add a backup carrier and revise the decision-making process.

As a result, the organization emerges from the crisis not merely having survived it, but more resilient than it was before.

This distinction matters.

The value of a strong supply chain professional lies not only in the fact that they once saved a particular shipment. Their decisions expand the company’s range of possible actions in the future.

Managers should know who these people are

There are not many professionals like this.

Most organizations know exactly who delivers savings, meets targets and maintains performance under normal conditions. Far fewer consciously assess another capability: the ability to generate executable solutions under pressure and with incomplete information.

Yet that capability is becoming critical in an era of global uncertainty.

Managers should pay attention to the people who start looking for an alternative while everyone else is still discussing the causes of the problem. To those who can connect procurement, logistics, production, finance and customer priorities in a single decision. To those who respond to “it is impossible” not emotionally, but with the next set of calculations and options.

These are the people companies should work hard to retain.

At the same time, they should not be turned into a permanent emergency service. Their experience has to remain inside the organization: translated into procedures, alternative routes, new agreements, supplier databases and tested scenarios.

As long as businesses are run by people, the human response to danger will not disappear. Some will fight, some will try to escape the risk, and some will freeze.

In supply chains, the last response is often the most dangerous.

Global uncertainty cannot be eliminated. No company can anticipate every conflict, closed strait, trade restriction, pandemic or next economic storm. But it can build a team of people who, when they find the road ahead closed, do not wait for someone to reopen it.

They start building another one.

About the author

Catherine Sharapova is a Supply Chain Strategist and Chartered Fellow (FCILT), holding the highest distinction from the Chartered Institute of Logistics and Transport. With over 15 years of experience and three degrees in International Business, Law, and Translation, she possesses a unique multidisciplinary command of the entire procurement cycle, from contract negotiation to customs clearance.

Currently driving strategy at Prometheus Group (industrial energy solutions), Catherine’s optimization of supply chains across China, India, and the UAE was instrumental in achieving a 4.2x increase in net profit in 2023. Her proprietary methodologies focus on mitigating payment infrastructure risks, accelerating inventory turnover, and reducing receivables.

She is a winner of a ‘Trailblazers’ at ‘Women in Supply Chain’ award by Food Logistics and Supply & Demand Chain Executive,  a ‘Rising Star’ in the ‘Pros to Know 2026’ by Supply & Demand Chain Executive, and a ‘Operational Transformer’ at ‘10 Women Shaping the Future of Logistics 2026’ by Global Trade Magazine. A member of the Procurement Union, she is a frequent speaker at major international forums. Her research on large-scale system development has been published in major trade and academic journals. Prometheus Group

Founded in 2010, Prometheus Group specializes in the supply of diverse battery types and raw materials for industrial production, mainly UPS. The company is a strategic partner for major explosion-proof equipment manufacturers, providing components and joint-production batteries utilized across various industrial sectors. With a robust operational footprint in China, India, the UAE, and the EU, the Group exhibits exceptional agility in supply chain management.

Today is Wednesday, Oct. 7

1. BD ties U.S. manufacturing expansion to tariff certainty

BD plans to invest $19 billion in U.S. operations, including $3 billion for manufacturing expansion, under an agreement offering conditional relief from future Section 232 tariffs on covered products and inputs.

BD  |   Read the original article

2. Record imports widen the U.S. trade deficit

The U.S. goods and services deficit increased 13.7% to $105.6 billion in August as imports reached $420.8 billion, with industrial supplies, capital goods and semiconductors contributing to the monthly increase.

U.S. Bureau of Economic Analysis  |   Read the original article

3. New Maryland shipyard targets submarine production constraints

Anduril and the U.S. Navy announced commitments totaling up to $6.6 billion for a software-defined Maryland shipyard intended to expand production capacity for components used in Virginia-class submarines

Anduril Industries  |   Read the original article

4. Energy producers push for alternate export corridors

Oil executives are urging investment in pipelines and ports that bypass vulnerable maritime chokepoints, arguing that recent disruption supports a shift from lean export networks toward more redundant energy infrastructure.

Reuters  |   Read the original article

5. Diesel and driver rules push trucking costs higher

Diesel prices and tighter driver availability are increasing U.S. trucking costs, with federal licensing and enforcement changes expected to remove substantial capacity and potentially support higher rates for carriers.

Axios  |  Read the original article

6. Global freight peak season stretches beyond expectations

Manufacturing demand, typhoon backlogs and Golden Week capacity reductions are extending peak-season pressure on some freight lanes, while returning Suez services are producing divergent rate and capacity conditions elsewhere.

DC Velocity  |   Read the original article

7. Warehouse automation supports later ordering cutoffs

First Supply says an Exotec Skypod system increased picking speed, concentrated more than 14,000 SKUs and enabled next-morning delivery for eligible orders placed by 5 p.m.

Robotics 247  |   Read the original article

8. Egg inventory signals begin moving toward balance

Cal-Maine sees early signs that U.S. egg supply and demand may be rebalancing as flock estimates, hatch rates and case production decline, although executives cannot predict when the glut will end.

Supply Chain Dive  |   Read the original article

Diesel markets remain vulnerable to geopolitical disruption, inventory shifts, changing refinery economics, and evolving freight demand. As transportation leaders build budgets and procurement strategies for the year ahead, understanding the forces shaping fuel costs has never been more important.

Join Breakthrough's market experts to hear the latest energy and freight market developments, the risks and opportunities shippers should be monitoring, and practical strategies to strengthen transportation decision-making in an uncertain environment.

Key Takeaways

• Understand the market forces driving diesel price volatility today
• Learn what current energy and freight market trends could mean for transportation budgets and procurement strategies
• Identify key risks and opportunities supply chain leaders should monitor in 2027
• Discover how market-based fuel reimbursement strategies can improve cost visibility and budget confidence

FEATURING:
Nick Skeen, Director, Business Development, Breakthrough and Matt Muenster, Chief Economist, Breakthrough
Moderator: Brian Straight, Editor in Chief, Supply Chain Management Review

Manufacturing procurement teams face mounting margin pressure as tariffs, input costs and supply disruptions make sourcing decisions more complex.

Join Bob Buterbaugh and Molly DeFilippo to explore how spend visibility and Advanced Sourcing Optimization (ASO) can uncover savings beyond price negotiations, evaluate cost and risk trade-offs, and accelerate decisions.

Learn how targeted cost takeout can help fund future procurement capabilities while protecting supply continuity.

What attendees will learn and take away

• Find savings beyond unit price by improving visibility across direct and indirect spend and prioritizing categories with actionable cost takeout opportunities.
• Use ASO to compare award scenarios that incorporate tariffs, freight, capacity constraints, supplier risk and sourcing alternatives.
• Balance savings with resilience by evaluating supplier viability, multi-sourcing and the cost of protecting supply continuity.
• Reduce decision time and manual effort through structured scenario analysis that helps constrained teams act as assumptions change.
• Build a self-funding approach by targeting near-term savings, measuring business impact and reinvesting value in future sourcing capabilities.

“I can unlock savings beyond price negotiations and use them to fund stronger sourcing capabilities, while making faster decisions that protect supply continuity.”

FEATURING:

Bob Buterbaugh, Account Manager, JAGGAER and Molly DeFilippo, Senior Business Consultant, BearingPoint

Moderated by Brian Straight, Editor in Chief, Supply Chain Management Review

Executive takeaways

In March 2026, a $240,000 copper shipment involving my company was recovered after sensor alerts showed the truck moving toward the wrong destination. Transport Topics later reported that the carrier that collected the load was apparently not the carrier expected to transport it.

That fact pattern reveals a risk hidden inside many modern supply chains. A company, driver, tractor, trailer, insurance policy, supplier, credential, or data record may appear legitimate when examined in isolation. The consequential question is whether the exact combination presented at the moment of execution was linked to the approved transaction and authorized by the right decision chain before custody, money, data, or control changed hands.

Most mature organizations know how to vet a third party. They review financial condition, capabilities, insurance, certifications, contracts, cybersecurity, and service history. But supply chains increasingly depend on third parties that perform or embed critical validation functions: brokers, screening platforms, insurance-verification services, auditors, supplier-risk providers, technology companies, and AI-supported decision systems.

That creates a second-order governance problem.

Vendor governance asks: Should we engage this provider?

Validation-event governance asks: Can we rely on this specific result for this specific decision?

A reputable validator can still produce a result that is incomplete, stale, based on the wrong source, linked to the wrong transaction, or weakened by an undocumented exception. When that happens, the enterprise can inherit confidence without actually possessing control.

Why “verified” is not enough

The word “verified” often compresses a complex process into a binary signal. It does not tell an operator what attribute was checked, which source was treated as authoritative, when the check occurred, whether material facts changed afterward, or who resolved conflicting information.

The gap becomes especially dangerous at handoffs. A supplier may have been approved months ago, but the bank account in today’s change request is new. A carrier may be legitimate, but the driver and equipment at the dock may not match the approved shipment record. An insurance certificate may be authentic, but the commodity or movement may fall outside the applicable coverage. An AI system may surface a plausible recommendation, but the underlying data may be incomplete or no longer current.

More validation layers can therefore create more confidence without creating more assurance.

Other risk disciplines already recognize the underlying principle. Federal banking guidance states that using a third party does not diminish an organization’s responsibility and calls for risk-proportionate due diligence, ongoing monitoring, meaningful performance measures, access to evidence, and audit rights. It also warns against incentives that reward processing volume or speed without adequate regard for accuracy and compliance.

NIST’s 2026 due-diligence guide for information and communications technology suppliers similarly organizes assessment around provenance, resilience, foundational practices, ownership or control, and supply-chain tiers. Neither source is a freight regulation. Both illustrate a mature governance principle: verification may be delegated, but accountability for the resulting decision is retained.

A practical test: VALIDATE the validation

Senior supply chain leaders can apply an eight-part test to any third-party validation that supports a consequential decision.

V - Validation scope. Define exactly what the validator is responsible for checking. “Carrier verified” is too broad. Entity legitimacy, operating authority, insurance applicability, driver identity, equipment identity, and shipment-specific authorization are different attributes.

A - Authoritative evidence. Establish which sources are acceptable and which source prevails when information conflicts. A screenshot, forwarded email, database record, insurer confirmation, government registry, and physical credential should not be treated as interchangeable evidence.

L - Linkage. Tie the result to the actual transaction, person, asset, credential, supplier, or shipment being presented. Validation of an organization does not automatically validate every actor or action taken in its name.

I - Interval. Define how long the result remains valid and which changes require revalidation. Time matters because authority, insurance, personnel, equipment, ownership, routing, and payment instructions can change after an initial check.

D - Discrepancy handling. Specify what happens when information is missing, inconsistent, or different from the approved record. A discrepancy should trigger a defined hold and escalation path, not an informal workaround created under schedule pressure.

A - Authority. Separate stop authority from release or override authority. Many people should be able to pause a questionable transaction. Far fewer should be able to clear it, and the rationale and decision owner should be documented.

T - Traceability. Preserve evidence showing what was checked, against which source, by whom, when, and why the final decision was made. A green indicator without recoverable evidence is difficult to audit, challenge, or improve.

E - Effectiveness. Measure more than completion rates. Track misses, false positives, unresolved exceptions, overrides, response time, downstream outcomes, and repeat failure modes. The validator's process should itself be validated.

Scale the control to the consequence

The answer is not to duplicate every check or impose maximum friction on every transaction. It is to match governance intensity to exposure.

Low-consequence decisions may rely on standardized screening. Elevated-risk decisions may require current evidence and explicit transaction linkage. High-consequence decisions may justify independent corroboration, non-discretionary holds, narrow override authority, and a complete audit trail.

The level of governance should follow the consequence of being wrong, not the convenience of the workflow.

The next control layer

Supply chains will rely on more validators, not fewer. Digital platforms, specialized service providers, automated screening, and AI will continue to expand the number of decisions supported by external assertions.

The risk is allowing “verified” to become a substitute for knowing what was actually validated.

Trust is not the absence of oversight. It is the product of clear scope, authoritative evidence, controlled exceptions, accountable decision rights, and measurable performance.

The next governance layer is not another badge or dashboard. It is the discipline to define, test, and audit the validation event before that event becomes authority to act.

About the author

Fernando Boom Aybar, CLTD, is Director of Transportation for Venture Metals International and Thalheimer Brothers. His work spans shipper-side transportation operations, carrier and broker governance, freight economics, cargo security, compliance, and technology-enabled transportation controls.

Today is Tuesday, Oct. 6

1. C.H. Robinson–RXO deal prompts a resilience review

C.H. Robinson’s proposed $5.8 billion acquisition of RXO would combine two major brokerage providers, giving shippers that use both firms reason to review provider independence, lane competition, contingency capacity and service baselines.

Logistics Management  |  Read the original article

2. Europe considers a faster defense against distorted trade

Germany and France are proposing a faster EU trade-defense mechanism that could restrict market access or cap sourcing from individual countries when subsidized imports threaten European manufacturing, although the plan remains under development.

The Wall Street Journal  | Read the original article

3. Schneider Electric moves to add PTC’s industrial software

Schneider Electric agreed to acquire PTC for approximately $22.6 billion, adding product-design and lifecycle-management software to its industrial AI portfolio in a transaction expected to close during 2027, pending approvals.

Robotics 247  |  Read the original article

4. Freight rates support European logistics earnings

European shipping and logistics companies are expected to report strong third-quarter results as elevated freight rates, resilient trade and disruption-driven demand support carriers and increase demand for customs and warehousing services.

Reuters  |   Read the original article

5. AI power constraints shift risk down the chip chain

Morgan Stanley says Nvidia and Broadcom are relatively insulated from data-center power constraints, while deployment delays could create greater risk for memory, optical-component and other suppliers deeper in the AI hardware chain.

Reuters  |   Read the original article

6. Britain reviews tariffs on Chinese electric vehicles

Britain is reviewing tariffs on Chinese electric vehicles as it seeks closer alignment with European industrial policy, but any measure would require evidence, legal process and consideration of retaliation and consumer costs.

The Guardian  |    Read the original article

7. RobCo’s expansion adds momentum to industrial robotics

Industrial-robotics company RobCo says its valuation surpassed $1 billion as it expands U.S. operations, prepares a 2027 commercial launch for its Alfie robot and provides liquidity to long-serving employees.

Robotics 247  |   Read the original article

8. Legacy systems remain a brake on enterprise AI

A vendor-sponsored survey of 945 technology leaders found legacy systems delayed or ended many AI projects, highlighting infrastructure, security and data-readiness constraints that can limit automation investments before deployment.

Express Computer  |   Read the original article

Executive takeaways

For a shipper using both C.H. Robinson and RXO, the proposed acquisition raises a question that goes beyond operating savings: How much does its transportation contingency plan depend on the two companies remaining independent?

If the transaction closes, two separately owned brokerage providers would become part of one corporate organization. That would reduce diversification at the ownership level for customers using both firms. Whether it weakens their resilience would depend on how the businesses are integrated, the carrier capacity behind those relationships and the other alternatives available on critical lanes.

The combination could also bring benefits. C.H. Robinson says the acquisition will improve network density and broaden its service offering, bringing together the companies’ truck brokerage and managed transportation operations, C.H. Robinson’s global forwarding business, and RXO’s expedited and last-mile capabilities.

There is also the AI-driven freight matching and potential cost savings, especially as C.H. Robinson drops its Lean AI onto the RXO portfolio. From a Robinson perspective, that may represent the greatest value to the transaction.

But for supply chain executives, the task is to evaluate the potential gains alongside changes in provider independence and competition. A larger network and a more diversified transportation strategy are different things, and this deal gives shippers a reason to examine both.

The companies announced Oct. 5 that C.H. Robinson has agreed to acquire RXO in a cash-and-stock transaction. The announced enterprise value is $5.8 billion and the implied equity value approximately $5.3 billion, according to a presentation prepared for investors. The companies expect to close in the first half of 2027, subject to regulatory and RXO shareholder approvals and other customary conditions.

C.H. Robinson expects approximately $300 million in net run-rate cost synergies within two years after closing, including improvements in cost-to-serve, operating efficiencies, shared services and third-party spending.

Revisit resilience and competition on overlapping lanes

If a shipper uses both C.H. Robinson and RXO, the proposed transaction raises an important question: How would common ownership change its provider strategy?

Two brokerage relationships would sit within one corporate organization after deal closing. For companies using the firms as independent alternatives, that would reduce brokerage and carrier diversification. If their contingency plans rely on separate organizations making capacity and service decisions, common ownership could weaken that aspect of resilience. Whether it does so in practice would depend on how operations are integrated and what other alternatives the shipper maintains.

There is also a potential pricing implication. On lanes where C.H. Robinson and RXO currently compete for the same shipper’s freight, the combination would bring two independent bidders under common ownership. If a shipper has few comparable alternatives, reduced competitive pressure could weaken its negotiating leverage and put upward pressure on the rates it pays.

The released materials related to the transaction do not establish lane-level competitive effects or predict higher rates. The outcome would also depend on alternative providers, operating efficiencies and carrier capacity.

The practical review starts with the shipper’s own bid history and routing guide: Where have the two firms competed, what share of freight do they handle, and which other providers can meet the same requirements?

Common ownership does not, by itself, establish that trucks or qualified capacity will disappear, but for a company that wishes to maintain independent brokerage relationships, it does matter.

For critical lanes, that review could help executives decide whether additional independent alternatives are warranted, even as they evaluate the combined company’s broader capabilities.

An operating model becomes part of the acquisition

C.H. Robinson is making its Lean AI operating model central to the transaction’s rationale. Its investor presentation describes more than 100 AI agents automating quote-to-cash tasks, supported by more than 450 engineers and data scientists. The company plans to apply its combination of lean processes, measurement and automation to RXO’s volumes.

“By applying our proven Lean AI model to RXO’s business, we expect to significantly enhance productivity to unlock compelling cost synergies,” Dave Bozeman, C.H. Robinson president and CEO, said in the announcement.

For customers evaluating the deal, it offers a starting point for a discussion around service outcomes. How will the projected savings translate to better service or lower rates?

For executives evaluating the combination, a useful baseline would include tender acceptance, on-time pickup and delivery, exception response, billing accuracy and the internal effort required to manage shipments. Quote speed and accuracy, and customer service levels will all be important metrics to watch as well.

The planned savings include shared-services efficiencies and other integration measures. Customers can ask how those changes would affect account support and exception response.

Network scale needs a lane-level test

The companies say the combination will improve network density and broaden their ability to support customers across modes and industries. They also expect the acquisition to expand C.H. Robinson’s proprietary datasets, improving its AI-driven sales, matching and procurement capabilities.

The presentation lists 75,000 shippers and 450,000 carriers for C.H. Robinson, compared with 18,000 shippers and 150,000 carriers for RXO. The presentation does not note carrier overlap, so those figures cannot be treated as a count of unique combined relationships.

For a supply chain executive, the useful analysis starts with the lanes, equipment and rates that matter to the business. The presentation noted improved network density, but that does not establish additional physical capacity. Where would the combination provide additional capacity? Will it mean there is more or less capacity during seasonal peaks? Which added capabilities address an existing operational need, and how would their value be measured?

Those are all questions shippers concerned by resilience should consider.

Make integration part of the service conversation

C.H. Robinson says it plans to integrate RXO primarily into its North American Surface Transportation division. Its 8-K identifies potential disruption to operations, personnel retention and integration challenges among the transaction’s risks. Those disclosures describe possibilities, rather than outcomes that have occurred. They also provide a starting point for shippers in ongoing discussions, both internally and externally.

For shippers using both companies, the proposed acquisition is a reason to revisit how their transportation contingency plans work. That review can identify where the combined network offers useful capabilities, and where independent alternatives remain important to service continuity and negotiating leverage.

Today is Monday, Oct. 5

1. G7 deploys emergency reserves as diesel prices climb

G7 nations agreed to release 100 million barrels of crude oil and refined fuels over the coming months, beginning with substantial diesel volumes as war-related disruptions and refinery constraints push transportation fuel prices higher.

Associated Press  |   Read the original article

2. Oil logistics keep delivered costs elevated

Middle East crude flows have recovered, but tanker rates, insurance costs, longer routes and limited refining capacity are keeping delivered oil prices elevated, underscoring logistics as the market’s continuing constraint.

Reuters  |   Read the original article

3. China’s LOGINK retreat reshapes the cargo-data contest

Reuters found China’s state-backed LOGINK cargo-data platform has retreated internationally amid stalled partnerships, unpaid bills and abandoned offices, marking a setback for Beijing’s effort to influence global shipping information infrastructure.

Reuters  |  Read the original article

4. Food costs rise as grain logistics tighten

The FAO food-price index reached 136.0 in September, up 1.5% from August, as Black Sea logistics constraints, weather risks and tighter crop supplies lifted cereal, vegetable-oil and sugar prices.

Food and Agriculture Organization of the United Nations  |   Read the original article

5. Factory orders post a modest August gain

U.S. factory orders rose 0.1% to $663.5 billion in August, while shipments were virtually unchanged, indicating limited demand growth as manufacturers balance investment against energy, trade and supply pressures.

U.S. Census Bureau  |   Read the original article

6. New standard defines mobile-robot user responsibilities

A new ANSI/A3 standard defines operating-phase responsibilities for industrial mobile robot users, including risk assessment, management of change, training and periodic reviews, completing a three-part framework for IMR safety.

Robotics 247  |   Read the original article

7. Humanoid partnership targets safer industrial deployment

Agility Robotics and FORT Robotics formalized work on Digit 5 safety, combining onboard and offboard controls, regulatory support and deployment engineering as humanoid robots move into more complex industrial environments.

Robotics 247  |  Read the original article

8. Warren CAT adds goods-to-person parts fulfillment

Warren CAT installed a Kardex-powered AutoStore system at its Midland, Texas, operation to manage parts orders with different urgency levels across its storefront, engine shop, warehouse and distribution activities.

Robotics 247  |   Read the original article

9. Automation forecast points to continued warehouse investment

A market-research forecast projects automated material-handling equipment sales reaching $61.14 billion by 2030, reflecting continued warehouse investment, although the commercial methodology and assumptions require editorial review.

Modern Materials Handling  |   Read the original article

Today is Friday, Oct. 2

1. Aircraft-part licensing becomes a U.S.-China pressure point

The Commerce Department has slowed export licensing for aircraft parts bound for China and is considering broader controls on landing gear and related components as leverage in continuing trade negotiations.

Reuters  |  Read the original article

2. Factory growth holds as input prices accelerate

U.S. manufacturing remained in expansion during September, but ISM’s input-price index jumped to 77.9 as energy costs and supply constraints intensified despite stronger new orders and factory hiring.

Reuters  |  Read the original article

3. CMA CGM completes FedEx Supply Chain acquisition

CMA CGM completed its $1.4 billion purchase of FedEx Supply Chain, adding 34 million square feet of warehousing and folding the contract-logistics operation into CEVA's expanding North American network.

Supply Chain 247  |   Read the original article

4. Ultium plans $1 billion Tennessee battery expansion

Ultium Cells plans to invest $1 billion in battery production and plant improvements at its Spring Hill, Tennessee, facility, adding 500 jobs and preparing new GM vehicle-cell capacity for 2028.

Supply Chain 247  |   Read the original article

5. Walmart adds oversized-goods fulfillment capacity

Walmart plans a $300 million Ohio fulfillment center for televisions, furniture and other oversized products, positioning inventory closer to customers and adding more than 300 jobs, although the construction timeline remains undisclosed.

Supply Chain Dive  |   Read the original article

6. FedEx places major electric-truck order

FedEx ordered 2,000 Harbinger electric trucks in a deal valued above $300 million, with the vehicles intended for pickup-and-delivery operations in the United States and Canada.

Supply Chain 247  |   Read the original article

7. SMB planners juggle lead-time and cost volatility

A Netstock survey found supplier lead-time swings, raw-material costs, freight expenses and demand shifts creating overlapping inventory-planning pressures rather than one dominant disruption for small and midsize businesses.

Supply Chain Dive  |  Read the original article

8. OSHA directs $10.2 million to safety training

OSHA awarded approximately $10.2 million to 67 organizations for hazard-control training and educational materials, emphasizing worker safety programs for small businesses and industries with elevated injury and fatality risks.

Work Safety 247  |   Read the original article

9. Avio begins $500 million rocket-motor factory

Avio USA broke ground on a $500 million solid rocket motor factory in Hurt, Virginia, adding domestic aerospace and defense manufacturing capacity as demand for propulsion systems increases.

Supply Chain 247  |  Read the original article

Executive takeaways

Gartner research shows that 84% of supply chain planning professionals expect spending on planning automation to increase over the next year. Yet many organizations remain unprepared for how that investment will change planners’ responsibilities. Only 49% of planners report being fully aware of the skills and competencies required for future career opportunities, while 73% of planning leaders identify the pace of change as a major barrier to performance.

This environment is accelerating the rise of the end-to-end planner, who looks beyond a single planning function to connect decisions across the supply chain and understand their wider business implications.

Preparing planners for this role requires more than teaching them to use new tools. Four capabilities will be particularly important as more planners step into this expanded role.

1. Master the art of problem framing

For years, planning organizations have invested heavily in developing problem-solving capabilities. Those capabilities remain important, but the greatest value increasingly comes earlier in the process.

As AI tools become more effective at generating forecasts, recommendations and scenarios, planners must determine which issues deserve attention in the first place. A demand fluctuation may appear to be a forecasting problem when it actually stems from a supplier constraint or a change in customer behavior. An end-to-end planner would bring together information from affected functions to identify the underlying issue before recommending a response. Framing the problem correctly shapes every decision that follows.

In an environment where AI can quickly generate answers, the ability to define the problem becomes a powerful differentiator. The value lies in bringing clarity to ambiguity and ensuring the organization is solving the right problem before it begins searching for solutions.

2. Be a decision architect

Planning decisions rarely affect just one outcome. A change that improves performance in one area can create unintended consequences elsewhere in the business.  As automation takes on more analytical work, planners will spend less time generating plans and more time shaping, governing and orchestrating the decisions that follow.

A decision architect makes those tradeoffs explicit. The planner clarifies decision ownership and aligns the response with operational realities and business priorities. This helps prevent one function from improving its own metric while creating additional cost or service risk somewhere else.

For example, a recommendation to increase inventory may protect product availability but introduce capacity or working capital concerns. The planner must clarify those implications, determine which stakeholders should participate and lead the organization to evaluate the decision against shared priorities.

This role increasingly involves working with AI-enabled forecasting, inventory optimization and scenario-planning tools. Planners need sufficient AI literacy to understand how a recommendation was generated and recognize when human review is needed. They must also be able to explain how a recommendation was evaluated and ultimately acted upon, particularly when the decision carries significant operational or financial consequences.

Effective planners combine strong business judgement and AI literacy to provide cognitive oversight of autonomous planning systems and orchestrate human workflows when needed.

3. Turn insight into enterprise execution

AI can generate insights, but insights rarely create value on their own. Value is created when organizations act on them. That means planners must become increasingly effective communicators and influencers. They need to translate analysis into business implications to help stakeholders understand tradeoffs and build alignment across functions that may have competing priorities.

A planning recommendation may require changes to production schedules, supplier commitments, transportation capacity or commercial plans. The end-to-end planner does not need to be an expert in every function, but must understand stakeholder priorities well enough to anticipate objections and establish a workable path forward.

This capability is becoming more important as supply chains grow more connected. Decisions are rarely confined to planning teams alone. They require coordination with manufacturing, procurement, logistics, commercial teams, and executive leadership.

4. Navigate uncertainty with confidence

The planner of the future must be comfortable operating in uncertainty. Supply chains are shaped by changing demand, supplier disruption, geopolitical events, technology shifts and evolving business priorities. Planners cannot rely solely on fixed processes or historical patterns; they must reassess assumptions, adjust plans and respond quickly as conditions change.

Adaptability does not mean reacting to every fluctuation. It requires recognizing which changes are temporary, which signal a broader shift and when intervention is necessary. Planners must balance speed with discipline while maintaining focus on business objectives, operational constraints, and customer commitments.

Risk tolerance is equally important. Not every risk can or should be eliminated, and decisions often involve choosing between competing exposures. A planner may need to determine whether to accept higher inventory to protect service, tolerate supply risk to manage cost or wait for better information.

By combining adaptability with informed risk judgment, planners can help organizations respond decisively while avoiding unnecessary disruption and overreaction.

The planner’s next chapter

Planning organizations now have to develop their talent with the same intention they bring to technology investment. Beyond access to AI tools, planners need opportunities to build judgment, understand the wider business and practice working across functional boundaries.

To make this possible, planning leaders will need to prioritize creating the time, experiences, and expectations that allow planners to grow into a broader role. AI may change how planning work gets done, but capable planners will determine whether better tools result in better decisions.

Gartner analysts are providing further analysis on this topic at the Gartner Supply Chain Planning Summit, taking place in Denver, CO on November 2-3.

About the author

Cristina Carvallo is a Senior Director Analyst with Gartner's Supply Chain practice where she helps clients develop and transform their supply chain planning processes and organization.

Today is Thursday, Oct. 1

1. Steel-producing nations prepare broader trade defenses

Twenty-eight economies agreed on a framework encouraging more anti-dumping, anti-subsidy and safeguard actions against excess steel capacity, while discouraging subsidies that keep uneconomic mills operating or support new overproduction.

Reuters  |   Read the original article

2. Caterpillar commits $1 billion to North Carolina expansion

Caterpillar plans to invest about $1 billion in North Carolina, including a new Sanford factory for compact track loaders and telehandlers, with the project expected to create more than 1,950 jobs.

Supply Chain 247  |  Read the original article

3. Supplier interruption cuts into Ford truck shipments

Ford said an unnamed supplier issue temporarily stopped F-150 production near Detroit and affected its Kansas City plant, reducing third-quarter wholesale shipments even though the automaker has restarted production.

Reuters  |   Read the original article

4. Nearshoring growth tightens cross-border trucking capacity

Executives said rising U.S.-Mexico freight demand is colliding with fewer qualified drivers, tougher customs enforcement, higher fuel costs and cargo-theft exposure, prompting some carriers to retreat from door-to-door service.

FreightWaves  |   Read the original article

5. FMCSA eases hours rules for fuel haulers

FMCSA granted interstate gasoline and diesel haulers a limited hours-of-service waiver through Dec. 16, citing global supply disruptions and higher seasonal demand from transportation and agricultural harvesting.

Transport Topics  |   Read the original article

6. Amazon and USPS test afternoon same-day handoffs

Amazon and USPS are piloting afternoon package induction for same-day delivery at two facilities, with a third location planned for Oct. 20 as the partners test tighter use of postal last-mile capacity.

Supply Chain Dive  |   Read the original article

7. Fake pickups dominate cargo-theft concerns

A SmartSense survey found 63% of security leaders reporting increased cargo theft, while 78% identified fraudulent pickups and carrier impersonation as leading concerns and 40% reported annual losses of at least $1 million.

Logistics Management  |  Read the original article

8. Field-safety data shows stubborn fatality risks

Field1st’s review of public safety data found U.S. workplace fatalities declined 4% in 2024, but transportation incidents, falls and contact events still accounted for nearly seven in 10 occupational deaths.

Work Safety 247  |   Read the original article

9. Motion-capture lab aims to improve humanoid training data

Innodata opened a New Jersey laboratory using optical motion-capture cameras to generate human-movement data for humanoid robot training and independently validate performance information produced by robotic systems.

Robotics 247  |  Read the original article

Executive takeaways

A customer accelerates a product launch. A supplier struggles to support the upside. Engineering qualification takes longer than expected. Inventory tightens, service escalations rise, and the leadership team asks the familiar question: Why did we not see this coming?

In many organizations, the honest answer is uncomfortable: We did. The customer had hinted at a change in timing. The supplier had expressed concern about incremental capacity. Engineering had raised readiness questions. A planner had noticed inventory exposure. The information existed, but it remained scattered across calls, emails, reviews, and functional teams. It never became enterprise intelligence.

This pattern reveals a limitation in the way supply chains have been transformed. Companies have invested heavily in enterprise resource planning, advanced planning systems, control towers, analytics, and automation. These capabilities have improved transactions, forecasting, visibility, and execution. Yet even sophisticated organizations still miss early signals because the signals often arrive as language, not data.

KEY CONCEPT

Supply chain surprises often occur not because information is missing, but because existing signals remain disconnected until options have narrowed. Also signals often arrive as language, not data.

From data scarcity to intelligence scarcity

For much of the modern supply chain era, the central problem was data scarcity. Information was incomplete, delayed, inconsistent, or trapped in local systems. The response was to standardize processes, integrate data, and create systems of record. That work remains essential.

The problem has now changed. Many companies have abundant data but insufficient intelligence. They have forecasts, scorecards, alerts, meeting transcripts, CRM notes, supplier reviews, customer updates, and project trackers. Still, leaders struggle to answer basic questions: Which supplier concern is becoming material? Which customer statement deserves action? Which internal assumption is no longer valid? Which escalation is noise, and which is an early warning?

This is intelligence scarcity: the inability to convert abundant information into timely, meaningful, and actionable insight. Visibility tells a company that a shipment is late or a supplier missed a commitment. Intelligence helps the company understand why the event matters, what related signals preceded it, and which decision should follow.

KEY CONCEPT

The challenge is no longer obtaining more data. It is converting abundant information into timely, connected, and actionable intelligence.

Managing intent

I call the missing discipline Managing Intent.

Intent is information about future behavior before action occurs. Customers express intent before placing orders. Suppliers express intent before changing commitments. Internal teams express intent before changing plans. Logistics partners express intent before service failures become measurable.

Intent is not the same as a forecast, commitment, or transaction. It is often incomplete and uncertain. A supplier saying that “upside may be difficult” has not missed a shipment. A customer saying that “executive attention is increasing” has not changed an order. An engineer questioning qualification timing has not delayed a launch. Yet each statement can narrow the range of plausible futures and provide valuable lead time.

The managerial challenge is not to react to every comment. It is to identify which signals matter, understand their context, test them against other evidence, and connect them to decisions. One isolated signal may be noise. Three related signals from a customer, supplier, and planner may constitute intelligence.

Managing Intent represents the next stage in supply chain evolution. Supply chains progressed from managing products, to managing information, to managing visibility. The emerging stage is understanding what stakeholders are preparing to do before actions become visible in formal systems.

KEY CONCEPT

Intent is information about future behavior before action occurs. Managing Intent turns early stakeholder signals into a disciplined management input.

A hierarchy of supply chain intelligence

The progression can be expressed as a five-level Supply Chain Intelligence Hierarchy. Each level adds a different management question. The higher levels do not replace the lower levels; they add context, anticipation, and ecosystem understanding.

KEY CONCEPT

Advantage shifts from understanding what happened to understanding what stakeholders are preparing to do and what the ecosystem collectively signals.

Where intent emerges

Intent travels through four collaboration networks. Each network offers a different view of the future. The greater opportunity is to connect the networks rather than interpret each one separately.

KEY CONCEPT

No single stakeholder holds the complete picture. Intent emerges across customer, supplier, internal, and ecosystem collaboration networks.

Intent Debt and Intelligence Latency

Two concepts make the problem measurable.

Intent Debt is the accumulated gap between what an organization collectively knows and what its formal systems recognize. Intent Debt exists when sales expects demand to rise but the forecast remains unchanged, when procurement hears a capacity warning but the supply plan assumes stability, or when engineering sees launch risk but the program remains green.

Like financial debt, Intent Debt compounds. A weak customer signal ignored today can become an allocation crisis later. A supplier warning left outside the planning process can become an expedite next month. The longer the gap persists, the fewer options leaders retain.

Intelligence Latency is the delay between the appearance of a signal and awareness by the person or forum able to act. The organization may not lack a signal; it may suffer from latency.

Reducing Intelligence Latency means routing validated signals to the right level of decision-making while options still exist.

KEY CONCEPT

Intent Debt is the knowledge formal systems have not absorbed. Intelligence Latency is the delay before that knowledge reaches a decision-maker.

Operationalizing intent through SCCI

Supply Chain Collaboration Intelligence, or SCCI, is the operating capability for Managing Intent. Its distinctive role is not merely to extract signals from individual conversations. Its role is to interlink those signals and reveal Collective Intent.

A customer may be preparing to accelerate a launch. A supplier may be protecting capacity because other customers are increasing demand. Engineering may be uncertain about qualification timing. Logistics may be seeing early congestion. Separately, each signal is ambiguous. Together, the signals reveal the direction in which the ecosystem is moving and the constraints that may prevent it from getting there.

Collective Intent is more than the sum of individual statements. It is the shared pattern created when customer priorities, supplier concerns, internal commitments, and ecosystem conditions are correlated across time. SCCI preserves the source and context of each signal, then connects signals by supplier, customer, product, program, region, and decision horizon. The result is a network view of intent rather than a collection of meeting summaries.

Consider a product launch where the customer is discussing an earlier ramp, sales expects upside, procurement hears that a component supplier is reserving capacity for another account, engineering has not closed qualification, and logistics sees longer transit variability. The collective pattern shows a widening mismatch between commercial aspiration and operational readiness. SCCI makes that mismatch visible and identifies the decisions needed to close the gap.

Collective Intent can also expose opportunity. Customer growth, available supplier capacity, and an underused production line may together support an earlier commitment or faster commercial response.

Capture uses approved collaboration sources. Understand identifies intent, risk, commitment, urgency, uncertainty, and decision relevance.

Correlate identifies reinforcing, conflicting, and dependent intent.

Recommend converts the collective pattern into evidence-backed options.

Learn compares signals, decisions, and outcomes so the capability improves over time.

SCCI is a management capability enabled by technology, not an autonomous decision-maker. Human judgment remains accountable for validating signals, weighing context, and approving action.

(Photo: Author)

SCCI operating model

Continuous feedback loop: outcomes refine signal definitions, confidence, and recommended actions.

KEY CONCEPT

SCCI interlinks customer, supplier, internal, and ecosystem intent to reveal Collective Intent: the emerging direction no single stakeholder can see alone.

A practical path to implementation

A practical starting point is a focused pilot. Select a high-value area such as strategic suppliers, new product launches, or constrained components. Define a signal taxonomy, including capacity risk, lead-time pressure, quality instability, cost pressure, launch readiness, and customer urgency. Review detected signals with business owners, record the action taken, and track whether the signal later materialized.

Leaders should measure outcomes rather than the number of signals collected. Useful measures include time from signal detection to validated awareness, the percentage of validated signals that lead to action, the proportion that later materialize, repeated-signal frequency, false positives, and the business impact of earlier intervention.

KEY CONCEPT

Begin with a focused pilot, validate signal quality, measure decision outcomes, and integrate proven signals into existing management processes.

How S&OP changes

Traditional S&OP was designed around a five-step planning cycle: Data Gathering, Demand Review, Supply Review, Integrated Reconciliation, and Executive S&OP. While highly effective at balancing supply and demand, the process largely begins after demand signals, supplier constraints, and business assumptions have already entered formal planning systems.

Intent-Driven S&OP introduces a new second step: Intent Review. Powered by Supply Chain Collaboration Intelligence, this step captures and interlinks customer intent, supplier intent, internal intent, and ecosystem intent to create a Collective Intent Brief before demand and supply plans are developed.

Rather than relying solely on historical data and forecasts, planners now incorporate stakeholder signals regarding future behavior, emerging opportunities, evolving constraints, and strategic priorities. This enables organizations to identify risks and opportunities earlier, evaluate scenarios more effectively, and make better-informed trade-off decisions.

Importantly, SCCI is not a standalone activity. It operates as a continuous intelligence layer across all six steps by capturing signals, understanding context, correlating patterns, recommending actions, and learning from outcomes. The result is an evolution of S&OP from a process focused on forecast reconciliation to a decision intelligence capability driven by Collective Intent.

(Photo: Author)

KEY CONCEPT

Intent-Driven S&OP adds an Intent Review step to the classical five-step process. By converting customer, supplier, internal, and ecosystem signals into Collective Intent, SCCI enables earlier visibility, better decisions, and more agile execution.

The future appears first in intent

Supply chains will continue to need better data, stronger planning, and improved visibility. But those capabilities alone will not eliminate surprise.

Customers signal before they order. Suppliers signal before they fail. Internal teams signal before plans change. Organizations that can hear, interlink, and act on those signals will preserve more options and respond with greater confidence.

The next frontier of supply chain transformation is not simply to forecast better. It is to understand Collective Intent earlier than others.

KEY CONCEPT

The next frontier is not simply better forecasting. It is understanding Collective Intent early enough to preserve choices and act with confidence.

About the author

Gautam H. Kamdar is vice president of TE Connectivity.

Today is Wednesday, Sept. 30

1. China warns of retaliation over proposed EU trade tool

China warned it could retaliate if the EU advances a trade instrument modeled on U.S. Section 301, raising risks of supply chain investigations, foreign-subsidy scrutiny and renewed friction for manufacturers sourcing across both markets.

Reuters  |   Read the original article

2. U.S. schedules final emergency oil release

The United States will offer up to 40 million barrels from its Strategic Petroleum Reserve, completing its 172-million-barrel contribution to coordinated releases while pressing European IEA members to fulfill their pledges.

Transport Topics  |   Read the original article

3. Battery shortage idles Stellantis assembly lines

Stellantis will pause production at two French assembly plants in late October because long-range battery supply from ACC has not kept pace with demand for Peugeot and Citroen electric vehicles.

Reuters  |  Read the original article

4. Panama Canal restores capacity and vessel draft

The Panama Canal Authority will raise Neopanamax transits to 10 daily on Oct. 15 and restore 49-foot drafts, bringing total daily capacity to 33 slots while warning unreserved vessels may still face delays.

FreightWaves  |  Read the original article

5. California raises cold-storage accountability

California enacted laws allowing contingency funds or insurance requirements for certain large cold-storage projects and higher penalties for serious safety violations, adding compliance and financial exposure following the Lineage warehouse fire.

FreightWaves  |  Read the original article

6. CBP tests advance truck export manifests

CBP will begin a two-year pilot with nine carriers on Oct. 23, testing advance electronic export manifests for U.S. truck freight moving to Canada and Mexico to improve targeting and reduce delays.

Transport Topics  |  Read the original article

7. Beef producers warn enforcement is chilling labor supply

Cattle groups from Texas, Kansas and Oklahoma warned that immigration enforcement is discouraging documented workers across feedyards, dairies, processors and related businesses, intensifying labor risk in an already constrained beef supply chain.

SupplyChainBrain  |  Read the original article

8. Bulk freight lead times stretch across key markets

Average bulk-freight lead times have stretched from 24 to 72 hours to seven to 10 days in tight Gulf Coast and central Mexico markets as driver constraints and selective carrier acceptance restrict capacity.

Trucking Dive  |   Read the original article

9. Manufacturing claims data puts worker engagement in focus

Travelers’ analysis of 1.2 million workers’ compensation claims found first-year employees accounted for 37% of claims, reinforcing calls for documented training, worker engagement and structured return-to-work programs across manufacturing operations.

Institute for Supply Management  |  Read the original article

Executive takeaways

For two decades, “AI in supply chain” has meant one thing above all: better forecasts. Statistical models gave way to machine learning, machine learning gave way to deep learning, and every generation promised the same outcome—sharper demand signals, tighter safety stock, fewer stockouts. That work mattered, and it is not finished. But it also quietly obscured a harder problem: forecasts, however accurate, do not run the supply chain. People do. And the moment a shipment misses a cutoff, a supplier short-ships an order, or a quality hold freezes a line, the AI’s job was traditionally done. A human took over, opened a ticket, pulled a report, made some calls, and fixed it.

That handoff point—the gap between insight and action—is where agentic AI is now making its case. Not as a better forecasting engine, but as a system that can sense a disruption, evaluate options against real constraints, and execute a resolution, often before a planner has finished reading the alert.

Two different jobs: predicting versus deciding

It’s worth being precise about what changes. Predictive and generative AI tools are fundamentally advisory. A forecasting model estimates demand; a GenAI copilot summarizes a disruption or drafts an email about it. Both stop at the recommendation. A human still has to notice the alert, weigh the trade-offs, and instruct a system to act—and in a global network generating thousands of exceptions a day, that human bandwidth is the actual bottleneck, not the quality of the underlying model.

Agentic AI closes that loop. An agent is given a goal, a set of tools (ERP transactions, TMS bookings, supplier portals, inventory systems), and a boundary of authority—and then it acts within that boundary without waiting for a person to approve each step. A pallet arrives damaged and the ASN doesn’t match the label data. Instead of sitting in a queue for a dock associate to investigate, an agent can cross-reference the shipment record, determine whether the discrepancy falls within tolerance, automatically trigger a partial receipt, flag the variance to the supplier’s system, and adjust the downstream allocation plan—all in the time it used to take a human to open the case.

Why exceptions, and why now

Exception handling is where this shift is landing first, for a simple reason: it’s where the “execution tax” is highest and most measurable. Every exception that sits idle in a queue, gets handed off late, or escalates past its resolution window costs money—in expedite fees, in rework, in inventory held hostage while someone figures out what happened. Those costs are predictable, repetitive, and rule-governed enough that an agent can be trusted with a meaningful slice of them, unlike, say, a genuinely novel strategic sourcing decision.

Three forces are converging to make this practical now rather than five years ago. First, the underlying data foundation has matured—cloud ERP, integrated planning suites, and real-time control towers now give agents a single, current view of orders, inventory, and logistics status to reason over, rather than yesterday’s batch extract. Second, large language models have become capable enough to serve as a reasoning layer that can interpret messy, unstructured signals—a supplier email, a customs delay notice, a weather advisory—and translate them into structured next steps. Third, and less discussed, is workforce pressure: with experienced planners retiring faster than they can be replaced, organizations are looking to encode that tribal knowledge into agents that can triage routine exceptions the way a twenty-year veteran would, freeing scarce human expertise for the judgment calls that actually need it.

What autonomy actually looks like in practice

The organizations doing this well are not handing agents a blank check. The pattern that’s emerging is a tiered governance model, roughly:

• Fully autonomous execution for high-volume, low-risk, well-understood exceptions—a minor quantity variance within tolerance, a routine reroute around a known carrier delay, a reorder trigger when inventory crosses a threshold.
• Automated recommendation with human approval for higher-stakes or ambiguous cases—a supplier substitution that affects cost or lead time, an allocation change that touches a strategic account.
• Human-led decisions with agent support for genuinely novel or high-impact situations—a facility outage, a geopolitical disruption, a multi-supplier failure—where the agent’s role is to assemble options, run quick scenario comparisons, and surface the trade-offs, but a person decides.

Early deployments bear this out. Procurement teams are using agents to request quotes from approved suppliers and rank responses without a buyer touching every RFQ. Category managers are letting agents handle supplier scoring and quote validation, stepping in only when the agent flags something outside normal parameters. Logistics teams are letting agents rebook freight automatically when a carrier misses a pickup window, rather than routing that decision through a person first. In each case, the agent isn’t replacing the planner’s judgment on strategy—it’s absorbing the high-volume, well-bounded decisions that used to consume the planner’s day.

The trust problem is the real bottleneck

None of this works if planners don’t trust it. A planner who has been burned once by a forecast they didn’t understand, or an automated action that made a bad situation worse, will simply route around the system—checking the agent’s work manually, which defeats the purpose. This makes two things non-negotiable for any agentic rollout: minimizing nuisance alerts so planners aren’t drowning in low-value notifications, and building in explainability—an audit trail showing what the agent saw, what it decided, and why—so a human can validate the pattern even when they don’t review every instance.

Governance has to scale with autonomy, too. As more decisions move to agents, and especially as multiple agents start coordinating across planning, procurement, and logistics domains, the attack surface and the blast radius of a bad decision both grow. That argues for clear authority boundaries per agent, monitoring for anomalous behavior, and—critically—humans who remain accountable for outcomes even when they didn’t click “approve” on any individual action.

What this means for supply chain leaders

The practical implication isn’t “replace your planners with agents.” It’s that the center of gravity in supply chain technology investment is moving from prediction to execution. A forecast that’s 2% more accurate is valuable; an exception that resolves itself in minutes instead of days is transformative, because it compounds across thousands of exceptions a week. Leaders evaluating where to start should look for the same profile every early success shares: high volume, well-understood rules, measurable cost of delay, and a clear boundary for what the agent is—and isn’t—allowed to decide on its own.

The forecasting era taught supply chains to see further ahead. The agentic era is teaching them to move faster once something goes wrong—and for most organizations, that second skill has been the more expensive gap all along.

About the authors

Souraj Roy is an MBA student in Supply Chain Management at Symbiosis Institute of Operations Management (SIOM), Nashik, with prior experience at Infosys. He holds SAP Certified Associate—SAP IBP for Supply Chain and Kinaxis Certified Maestro Author Level 1 certifications.

Samriddha Basu, CAPM®, is an MBA student in Operations & Supply Chain Management, currently interning at CEAT and serving on the Consulting Committee. He holds the Certified Associate in Project Management (CAPM), SAP Certified Associate—SAP IBP for Supply Chain, and is a Kinaxis Certified Maestro Author Level 1.

Today is Tuesday, Sept. 29

1. Germany moves to block Cosco’s logistics acquisition

Germany reportedly plans to block Cosco’s proposed 80% acquisition of Hamburg logistics company Zippel, citing concerns that control of container movements and related software could create strategic dependencies and expose sensitive supply chain information.

Reuters  |   Read the original article

2. Tariff proposal opens limited U.S.-China relief

A bilateral trade board recommended preferential tariff treatment for $30 billion of non-sensitive goods in each direction, but the United States and China have not specified reduction levels or an implementation timetable.

Supply Chain Dive  |   Read the original article

3. Suez routing trims India-to-Savannah transit

Maersk and CMA CGM are returning selected India-Savannah services to the Suez Canal, cutting transit times by as much as 14 days and potentially reducing importers’ inventory-carrying and vessel-capacity requirements.

FreightWaves  |   Read the original article

4. Shippers seek longer pause on China-linked vessel fees

Retail, shipping and logistics groups asked USTR to extend the suspension of fees on China-linked vessels beyond Nov. 9, warning that renewed charges would compound elevated transportation costs and network uncertainty.

Supply Chain Dive  |   Read the original article

5. AAR deal expands constrained aircraft-maintenance capacity

AAR agreed to acquire a 65% stake in MRO Holdings for about $1.8 billion, expanding aircraft-maintenance capacity as airlines contend with servicing demand and continuing delays in new-jet deliveries.

Reuters  |   Read the original article

6. Puma turns idle warehouse capacity into shared infrastructure

Puma is allowing Maersk to offer unused capacity in three automated U.S. distribution centers to other customers, converting 2.3 million square feet of logistics infrastructure into a multi-client fulfillment network.

FreightWaves  |  Read the original article

7. Technology sourcing strategies add exit plans

Procurement leaders are being urged to prequalify alternate technology providers, improve data portability and fund executable exit plans as geopolitical scrutiny and vendor concentration increase operational and compliance risks.

SupplyChainBrain  |   Read the original article

8. LTL shipment trends split across major carriers

August shipment trends diverged among major LTL carriers, with XPO and Saia reporting gains while ArcBest and Old Dominion declined, reflecting different customer strategies and exposure to truckload spillover freight.

Supply Chain Dive  |   Read the original article

9. Cargo-theft survey points to more sophisticated attacks

A SmartSense survey of 150 loss-prevention professionals found 63% reporting increased cargo theft and 81% seeing more sophisticated attacks, with fraudulent pickups, carrier impersonation and GPS interference among leading concerns.

SupplyChainBrain  |   Read the original article

Executive takeaways

The most useful conversation at a conference does not always begin with a microphone. It might start over breakfast with someone wrestling with the same planning problem, continue during a break with a speaker who has lived through an implementation, or take shape at a reception after the day’s sessions are done.

The 2026 NextGen Supply Chain Conference is creating time for those conversations alongside its program of keynotes, panels and small group discussions. And it’s doing with a truly Nashville flair. Attendees gathering at the W Nashville hotel from Oct. 21-23, 2026, will have two evening receptions, a networking lunch and live performances by three Nashville musicians.

An opening night to meet the NextGen community

The conference begins Wednesday, Oct. 21, with a welcome reception and networking from 5 to 7 p.m. in the W Nashville’s Living Room. It is an opportunity to meet fellow attendees, speakers and industry partners before the main program begins Thursday morning.

Nashville-based singer-songwriter Nick DeLeo will perform at the reception. Originally from Cape Cod, Massachusetts, DeLeo has made Nashville home since 2013. His music draws on country, rock and blues, combining original songs with his interpretations of familiar favorites.

[Click to view the NextGen agenda]

A Thursday built for conversation

Thursday, Oct. 22, opens with breakfast from 7:30 to 8:30 a.m. in the Great Room Foyer and Terrace, sponsored by Verity. Attendees can bring their coffee and breakfast into the general session, which begins with the NextGen Supply Chain Awards. Zion Solutions Group sponsors the awards program.

The morning break, from 10:15 to 10:45 a.m. in the Great Room Foyer and Terrace, is sponsored by Corvus Robotics. It offers another chance to compare notes before the Small Group Sessions begin. Those discussions bring practitioners and technology providers together around real implementations, with sessions repeated in the afternoon so attendees can join more of the conversations relevant to their work.

At 12:15 p.m., attendees will gather at Zaytinya for lunch and networking, sponsored by Gather AI. Independent Nashville singer-songwriter Emma White will perform during lunch. White blends country storytelling with pop and R&B influences and runs her own label, Whitehouse Records. Rolling Stone has included her among its “10 New Country Artists You Need to Know.”

After the afternoon program and a 3:10 to 3:30 p.m. break, Thursday concludes with a rooftop reception from 5 to 7 p.m. at the W Nashville’s Butterfly. Performer Scooter Carusoe, the songwriting name of Travis Hill, has worked in Nashville music for two decades and co-founded the independent music publisher Carnival Music. His songs have been recorded by artists including Kenny Chesney, Darius Rucker and Brett Eldredge. The rooftop gathering gives attendees time to continue conversations from the day’s sessions while hearing the work of a songwriter closely tied to the city.

One more morning to make connections

The conference continues Friday, Oct. 23, with breakfast from 7:30 to 8:30 a.m. in the Great Room Foyer and Terrace. Attendees can take breakfast into the morning general session, which includes the solution provider awards, sponsored by Zion Solutions Group, and presentations and discussions on AI, supply chain planning and talent. A morning break runs from 10:15 to 10:35 a.m.

[Click to view the NextGen agenda]

NextGen’s theme, Innovate. Upskill. Transform., runs through the formal agenda. The meals, breaks and receptions create room for a different kind of learning: asking a follow-up question, finding a peer facing a similar challenge and building a relationship that lasts beyond Nashville.

The 2026 NextGen Supply Chain Conference takes place Oct. 21–23 at the W Nashville. View the conference agenda and register to attend. Attendees seeking the conference hotel rate should book by Wednesday, Sept. 30, at 5 p.m. Eastern, subject to availability.

I. The Better Tanks Lost

In May 1940, the French army held what looked like a decisive advantage. It fielded more tanks than the German force massing on its border, and in the measures that armies cared about the most—armor thickness and gun caliber—several of its models were superior to anything the Germans could put in the field. By the cold arithmetic of hardware, France should have held.

Six weeks later, France had fallen.

The explanation was not in the steel. It was in everything around the steel. French doctrine treated the tank as an infantry support weapon, parceled out in small groups along a wide front, advancing at the pace of the foot soldiers beside it. Command was centralized and methodical. Orders flowed down from the top, and a unit that saw an opportunity in front of it waited for permission to take it. Many French tanks carried no radio at all, so their crews could not coordinate at speed even when they wanted to.

The German edge was not a machine but a system. Radio-equipped panzers let commanders concentrate force wherever a gap appeared. Junior officers were trusted to act on what they saw without waiting for headquarters, a command philosophy the Germans called Auftragstaktik, or mission command: leaders were given intent and left to work out the how for themselves. And Stuka dive-bombers, working as flying artillery in direct support of the armor, struck strongpoints and severed communications ahead of the advance. Tank, aircraft, and radio reinforced one another into a single fast-moving whole. Rather than dispersing tanks across the front in support of infantry, as the French did, the Germans concentrated their tanks into an armored “fist” designed to punch holes through enemy lines. 

The French could have matched any one of those pieces, and on armor they largely outmatched the Germans. They still lost, because what defeated them was not a better weapon, but an organization built to find an opening and exploit it before they could respond.  The German plan carried enormous risk, and the campaign had its share of luck, but that is the point: the German system was built to exploit luck when it arrived, and the French system could not exploit anything at all.

France had the better arsenal. Germany had the better doctrine and the better formation, and it was not close.

We open with this paradox because it is the most precise statement we know of the problem facing supply chain organizations today. The lesson is not that hardware does not matter. It is that hardware alone does not save you. A technology, however advanced, does not generate the leadership and the organization required to use it. Those must be built deliberately, and most organizations never build the column. They buy the tanks and keep the doctrine.

That is the trap this article is about. Artificial intelligence, and Agentic AI in particular, is arriving as the most capable arsenal supply chains have ever been handed. The organizations that win the next decade will not be the ones that buy the most of it. They will be the ones willing to change how they lead, and how they are organized, to let it work.

II. The world has changed, and it is not changing back

The France analogy would be only a history lesson if the ground beneath supply chains were not actually shifting. It is. Two forces have turned the old way of running a supply chain, plan carefully and then execute the plan, into a liability.

1. Geopolitical fragmentation. Friend-shoring, tariff escalation, export controls, and the hardening of regional trading blocs have turned political alignment into a sourcing variable. Where a company can buy, build, and sell is now shaped as much by diplomacy as by cost. This is not a storm to wait out. It is the climate.
2. Economic volatility that defeats forecasting. Demand signals, input costs, and lead times move faster and further than planning cycles can absorb. You cannot forecast your way through this. You can only out-adapt it, which shifts the advantage from those who predict best to those who adjust fastest.

Onto this unstable ground arrives the technology itself. When Gartner asked global supply chain leaders to name the force most redefining their work, AI and Agentic AI ranked first. Ahead of the use of economic policy as a political weapon. Ahead of deglobalization. Ahead of political instability. The people running these organizations are telling us that the largest change is not the geopolitics they read about every morning. It is the capability landing on their own desks.

There is a trap hidden in that ranking. Faced with a dominant new technology in a volatile world, the instinct is to treat the problem as one of acquisition: buy the capability, and survival follows. It does not. The organizations that endured past inflection points were not the ones with the most advanced tools. They were the ones built to change how they worked when the tools demanded it.

Consider two incumbents who watched the same wave arrive. Best Buy and Circuit City both had the internet, and both saw Amazon coming. Circuit City answered with a spreadsheet. In 2007, while still profitable, it fired some 3,400 of its most experienced salespeople for the offense of having become its highest paid, and replaced them with cheaper, greener hires. Best Buy rebuilt the work instead. It invested in its floor staff and Geek Squad, matched online prices, and turned its stores into shipping points to fulfill online orders from the shelf. By 2009, Circuit City was liquidated. The layoffs were not the only cause, but they were the symptom: a company that met a transformative threat by cutting the very people who knew its customers. Best Buy is still here. Same tool, same threat. One company rebuilt how it worked, and the other tried to cut its way to the future.

If this pattern feels familiar, it should. The supply chain field has run this experiment before, and more than once. Material Requirements Planning (MRP) arrived in the 1970s as the system that would finally synchronize the factory, and an entire certification industry grew up around it. Computer Integrated Manufacturing followed in the late 1980s, then Total Quality Management (TQM), followed by Lean, then the Theory of Constraints. Each was hailed as the decisive advance, and each underdelivered against the promise for the same reason. When the results disappointed, the field blamed the data, the software, or the implementation, and almost never looked at the two things that actually governed the outcome: how decisions were made, and how the organization was arranged to act on them.

The most instructive case is Lean. The tools of the Toyota Production System are public, documented, and freely copied. What cannot be copied off a shelf is the management system and the culture that Toyota built around them. One of us spent years studying that system from the inside, and watched a major automaker construct a faithful replica of a Toyota plant, down to the layout and the andon cords, and fail with it. They had the trappings without the essence. Anyone can buy the tools. What makes the tools work is an organization that has reorganized itself around them. AI is the newest tool on the same old shelf, but the same rule applies.

Adaptation, in other words, is not a technology problem; it is an organizational design problem. And like any design problem, it has a structure.

III. The three pillars: Arsenal, Doctrine, Formation

If adaptation is an organizational problem rather than a technological one, then it has a structure, and that structure has three parts. We call them the Arsenal, the Doctrine, and the Formation (borrowing the language of 1940 because the failure is the same). The Arsenal is what the supply chain can do. The Doctrine is how its leaders decide. The Formation is how the organization is built to turn those decisions into action. The central claim of this article is easy to state and hard to live by: organizations fail at new technology because they invest in the first pillar and neglect the other two.

The capability-first trap

Most organizations approach transformation by acquiring capability and assuming the rest will follow. They fund the platform, stand up the analytics team, and launch the pilots. The capability arrives. The transformation does not. The reason is visible in the economics of where AI value actually comes from. By BCG’s accounting, roughly 10% of the return on an AI initiative comes from the algorithm itself, and 20% from the technology and data plumbing around it. The remaining 70% comes from people and process: literacy, leadership alignment, change management, and a credible path to proving value. The Arsenal, in other words, is the smallest slice of the pie.

And yet organizations spend as if the opposite were true, pouring the budget into the algorithm and the plumbing and treating the people and the process as a rounding error. Randstad’s 2026 Workmonitor, drawn from more than 27,000 workers and 1,200 employers across 35 markets, found that nearly two-thirds of employers invested in AI in the past year, while across most sectors that capital spend is outpacing the investment in the skills needed to use it. Randstad is blunt about the result, calling it an AI reality gap that is a skills problem, not a technology one.

Leaders are not blind to it. In Mercer’s 2026 talent study, 63% of executives called redesigning work around AI their highest-return investment, while only 32% said their workforce was ready for it. The knowing is not the problem; the spending is.

There is a deeper reason the spending runs ahead of the work, and it sits in the C-suite. For many leaders, the responsibility feels discharged the moment the check is signed. The budget is approved, the vendor is selected, the announcement goes out, and the problem is now someone else’s to make real. But writing the check is the easy part, and ultimately the cheapest part. The expensive, uncomfortable work is everything that comes after: rewiring how decisions are made, redrawing who owns what, and resetting the incentives that govern daily behavior. Leaders who believe their job ends at the check have, without realizing it, become the constraint.  It is the wrong message, and everyone below hears it clearly. Writing the check is the easy part. The expensive, uncomfortable work is everything that comes after.

Four verbs, three pillars: why all or none

A future-proof organization does four things quickly and well. It assesses what is happening and what it means. It responds to near-term shifts in demand and supply. It adapts its design to deeper structural change. And it evolves its own capabilities over time. Those four verbs are the visible signature of an organization built for this environment, and every one of them draws on all three pillars at once. A deficit in any single pillar quietly drains the other two. Assessing leans on the Arsenal, but it dies if leaders filter out the signals they do not want to see. Responding and adapting depend on capability, but they require leaders empowered to act and a structure that lets them. Evolving is almost entirely a matter of Doctrine and Formation. Because each verb reaches across all three pillars, no single pillar can future-proof anything. That is the whole of the argument. The rest of this two-part article is its demonstration.

IV. The curse of knowing

If the problem is organizational, the next question is the hard one. Why do capable organizations, run by intelligent people, so reliably fail to make the change they can see coming? The answer is a particular kind of trap, because it is the central obstacle to future-proofing. Call it the curse of knowing.

Managers rise through an organization because they are good at achieving certain outcomes, such as cost reduction, lead time reduction, or throughput improvement.  They make sound decisions, and when the next decision arrives, they reach for the judgment that worked before. As long as the environment holds steady, that instinct is exactly right. It is what experience is for. The curse appears only when the environment shifts underneath them. That is because these same skills often do not translate well to the new environment and the new outcomes.  The very pattern-matching that produced a career of good calls now produces confident, well-reasoned, wrong ones. An organization that keeps executing what made it successful can become the best in the world at running a process that no longer matters.

Clayton Christensen spent a career showing why this failure is rational. The managers who miss a transformative technology are usually running their resource allocation process exactly as designed: it funds what serves today's customers and today's metrics, and it starves whatever does not. The curse lives in the machinery that decides where money and talent go, not merely in individual judgment, which is why speeches alone never break it.

This is why the most dangerous sentence in any organization is “This is not how we did it in my day.” The honest reply is that it is not your day. It is a different day, and the question is no longer whether the old approach was good. It is whether it still fits the world in front of you.

What makes the curse so hard to break is that the old approach is not merely a habit. It is a record of success, and success is adhesive. We made a version of this argument in a previous article in these pages, tracing how organizations reject transformative innovations for organizational rather than technical reasons, from Kodak and digital photography, to Xerox and the personal computer, to an army that clung to single-shot rifles long after the repeater existed. The mechanism is consistent. Past success builds mental models that quietly filter out the signals that do not fit, so the very efficiency that makes a culture strong at optimization makes it weak at transformation. In that article, we organized readiness into three dimensions, whether an organization can Observe change, Accept it, and Act on it, and the curse of knowing is where the first two break down: the signal arrives, and the mental model filters it out. The sharpest illustration, though, is not from business at all.

The way out is a single discipline. When an uncomfortable idea arrives, do not ask only whether it fits your current model. Ask what would have to be true for it to be right, and then go find out. That question is the antidote to the curse of knowing, and the rest of this two-part article is about building an organization that asks it by default.

V. The metamorphosis of technology

The tank is a useful teacher precisely because not every technology is a tank. Most are not. Before deciding how hard to change, a leader has to be honest about what kind of technology has actually arrived, because any technology stands in one of three relationships to the work.

• A complementary technology makes the existing way of working a little better. The stakes are low, and the doctrine is left untouched.
• A substitutive technology replaces a task or a tool inside the current process. The savings can be real, but the shape of the organization does not have to change to capture them.
• A transformative technology is different in kind. It changes what is possible, and its value cannot be captured by bolting it onto the old process. Capturing it requires leaders to free themselves from constraints that are no longer there. That is the level that demands a new doctrine and a new formation, and it is, by its nature, strategic.

Agentic AI sits at that third level. It does not merely advise. It senses, reasons, plans, and acts within boundaries, and in doing so it collapses entire layers of decision-making that used to be the work itself. Consider what that means in practice. In our interviews, assembling a single sourcing business case meant chasing data from five people in five departments, and by the time the case came together, the opportunity had often moved on. An agent handed the same task pulls the spend data, cost models, and demand picture in minutes and drafts the case itself, so the decision that once waited weeks on assembly now waits only on judgment. The friction this creates is not technical. It is that the machine runs at machine speed while the approval chain around it still runs at calendar speed. A company that treats such capability as a faster tool, and leaves its doctrine and formation intact, has bought a tank and kept the marching orders.

The deepest reason Germany conquered France is the same reason this matters now. Col. John Boyd studied campaigns like it and distilled the lesson into the OODA Loop: the side that can Observe, Orient, Decide, and Act faster gets inside the other side’s loop. Auftragstaktik was a faster decision loop. Boyd’s larger lesson is that rapid change can only be met with rapid reinvention, through iterative cycles of learning tested against the real world. Future-proofing is the ability to cycle that loop faster than the environment around your changes.

This is the deeper reason doctrine and formation matter, and it is easy to miss. We tend to treat them as the means by which an organization absorbs a new technology. They are more than that. Their real work is to raise an organization’s capacity to notice the unexpected and act on it faster than rivals can, which is exactly what Boyd’s loop describes. The side that observes, orients, decides, and acts more quickly does not merely move faster. It shapes a situation the other side is still trying to understand. Technology supplies the speed. Doctrine and formation decide whether the organization can see the opening that speed creates.

Figure 1. Boyd’s OODA Loop. The Arsenal supplies the speed of each cycle. The Doctrine and the Formation determine whether the organization sees, and can exploit, the openings that speed creates.

For a real, current demonstration of this principle in action, look at how Ukraine has used inexpensive drones against a far larger and better-funded Russian military. The technology itself is not exotic, and it is available to both sides. What Ukraine added was the other two pillars. It pushed authority to the operators at the front, who configure and adapt their systems rather than waiting for a finished product to arrive. It built an incentive structure that rewards operators for hitting the targets that matter most, feeding real battlefield demand straight back into the next generation of design. And it placed the effort under leadership that had already digitized the machinery of the state, so that the technology, the decision rights, and the organization were aligned from the start. Drone and counter-drone tactics there now turn over in weeks, while the doctrine and procurement cycles of larger militaries still move in years. Dr. Oleksandra Molloy of the Australian Army Research Centre, who studied that adaptation through Boyd’s own OODA lens, concluded that the decisive edge lies not in any single aircraft but in the human and institutional capacity to keep adapting under fire. It is France 1940 run in reverse, and it is the clearest modern proof that alignment, not hardware, is the multiplier.

VI. Where supply chain organizations actually sit

All of this resolves into a single picture (see below), and it is the one thing we want you to take away from this article. Place your organization on two axes. The horizontal axis is the Arsenal: how much technology and automation capability it has, from low to high. The vertical axis is the alignment of the Doctrine and the Formation: how adaptable its leadership and its structure are, from low to high. We plot them together deliberately, because the vertical axis measures their alignment, not their sum. A gardener's doctrine trapped inside a chess master's structure scores low on this axis by definition, and so does the reverse, because a decision philosophy and a structure that contradict each other produce frustration rather than speed. Four quadrants follow.

The punchline of the matrix is uncomfortable. Spending on the Arsenal moves you to the right, not up. It can carry a Fragile company straight into Automated but Brittle, which feels like progress and is not. Only work on the Doctrine and the Formation moves you up, toward Future-Proof. No check written to a vendor will do it.

The numbers describe the Automated but Brittle box precisely. The spending is real and accelerating. Gartner projects that spending on supply chain management software with Agentic AI capabilities will grow from under $2 billion in 2025 to $53 billion by 2030. What has not kept pace is alignment. Even among supply chain organizations already deploying AI, only 23% report having a formal AI strategy to direct it. And actual implementation remains thin: as of 2025, just 5% of companies using supply chain software had Agentic AI features in place, a figure Gartner expects to reach 60% by 2030. That is the signature of high spending meeting low alignment. The money is real. The movement up the vertical axis is not.

Placing yourself on this matrix honestly is harder than it sounds, and most organizations fail the test. When we asked supply chain teams where they sat on the path from manual work to full autonomy, nearly all of them placed themselves near the top, comfortably automated. Pressed on specifics, most turned out to be one or two rungs up, using AI as a smart search engine rather than a system that decides and acts. The gap between the self-image and the reality is not vanity. It is dangerous, because an organization that misjudges where it stands will misdiagnose what it needs, conclude that it is merely behind on technology, and write another check for the Arsenal when the real deficit is in the Doctrine and the Formation. Two questions have to be answered honestly before any of this can proceed. The first is where you actually are. The second is why you are buying at all: to amplify basics you have already mastered, or to compensate for basics you have not. Technology bought to amplify strength compounds it. Technology bought to paper over weakness only automates the weakness.

When we looked at where the organizations in our research actually cluster, the pattern held. The capability was rarely the thing missing. The alignment was.

VII. What the field is telling us

None of this is theory. Researchers from the University of Tennessee-Knoxville spoke with more than 30 organizations, split between the solution providers building these systems and the supply chain teams trying to put them to work, across industries from consumer goods to homebuilding supply to cold storage. The same few patterns surfaced again and again, and every one of them points away from the technology.

The first is that the binding constraint is almost never the model. Firms race into AI on dirty data, crippling their ability to draw actionable insights. One senior principal at a major consultancy named the failure mode exactly: told to build data lakes, companies “tended to build Lake Superior versus Lake of the Ozarks,” vast where they needed only the water the decision in front of them required.

The second pattern is pilot purgatory. Organizations run pilot after pilot with no mechanism to scale them, vague criteria for success, and no decision timeline. The capability is usually real. What is missing is any way to convert it into operating reality, which is a Formation problem, not a technology one.

Underneath these sat a consistent set of visibility gaps: procurement run in isolation, legacy systems that resist integration, immature governance, thin third-party data, blindness past the first tier of suppliers, compliance handled as a recurring fire drill, and, most telling of all, a why gap. Organizations reliably record what was decided and almost never record why. That is precisely the context debt that makes an automated decision impossible to audit when it later matters.

The why gap is more dangerous than it looks, because the organizations creating it are often dismantling the only record of the why at the same moment. As experienced people are laid off, decades of undocumented judgment, the reasoning behind decisions that were never written down, walk out the door with them. This is the parable of Chesterton’s fence applied to the enterprise: before you tear a fence down, you had better know why it was put up. An AI trained on what an organization decided, but never on why it decided, or on the alternatives it considered and deliberately rejected, will confidently optimize toward the wrong objective. One of us saw this directly in transportation pricing, where the profit-maximizing option was sometimes passed over on purpose, because a different option strengthened the competitive position or pressured a rival more than it cost the firm. An automated system trained only on the chosen outcomes, blind to the strategic reasoning behind them, would learn to chase a number that was never the real goal. The judgment was the asset. The decision was only its residue.

The most revealing answer came when we asked one supply chain leader what he would change if he were handed a magic wand. He did not reach for cleaner data or better systems. He said he would put every employee through a course on digital logic. Asked to name the single highest-leverage fix available to him, an experienced operator chose human development over technology. That is the thesis of this article, stated by the field itself.

VIII. The question this raises

The diagnosis is now on the table. The world has shifted in ways that reward adaptation over prediction. The technology arriving is transformative, not incremental. The organizations that fail with it fail for organizational reasons, not technical ones. And a single matrix locates any company on two axes, with most heavy spenders clustered in the Automated but Brittle quadrant, having bought the Arsenal and left the Doctrine and the Formation untouched.

This points to an asymmetry worth naming before we turn to the work itself. A competitor can fast-follow the Arsenal. The same models, platforms, and tools are available to anyone willing to write the check. The vendors are not selling competitive advantage; they are selling, in the case of AI, licenses and models. What a competitor cannot fast-follow is the Doctrine and the Formation, the way an organization decides and the way it is built to act, because those are earned through sustained investment rather than purchased. And there is a third thing, harder still to copy: the learned ability to exploit the openings that doctrine and formation create. That capacity is built over years of acting, missing, and adjusting, and it cannot be acquired on the timeline of a procurement cycle. It may be the most durable advantage of all.

Knowing where you sit, however, is not the same as knowing what to do about it. That is the subject of Part Two: what are the moves that carry an organization upward, toward Future-Proof?  What must change in how leaders lead, in how the organization is built, and in how its people are developed to question and govern the machine? And how does a leadership team know, honestly, whether it is ready to begin?

About the authors

Steven A. Melnyk, Ph.D., is an emeritum professor of Supply Chain Management at Michigan State University. He has co-authored 25 books, more than 100 refereed journal articles, and more than 300 practitioner articles. He is a past member of the APICS Board of Directors. Since retiring in 2024, Dr. Melnyk is now part of the Executive Development Program at Michigan State University and can be reached at Melnyk@broad.msu.edu

Alan Amling, Ph.D., is an assistant professor of practice in Supply Chain Management at the University of Tennessee’s Haslam College of Business and CEO of Thrive and Advance, LLC. This article is adapted from “Organizational Velocity: Turbocharge Your Business to Stay Ahead of the Curve.” He can be reached at aamling@utk.edu.

I. The story so far

To set the stage (from Part One): In May 1940, France fielded more tanks than Germany, and several of its models were better armored and better armed. Six weeks later, France had fallen. The cause was not the steel. France parceled its tanks out along a wide front, kept command centralized, and fitted many of them with no radio at all. Germany put radios in its panzers, concentrated its armor, and pushed decision authority down to the officers who could actually see the battlefield. It tied tank, radio, and dive-bomber into one fast-moving system, able to find an opening and exploit it before the French command could respond. France had the better arsenal. Germany had the better doctrine and the better formation.

That is the precise shape of the problem facing supply chains as Agentic AI arrives. Every transformation rests on three pillars. The Arsenal is the capability, what the supply chain can do. The Doctrine is the management style, how its leaders decide. The Formation is the organization, how the enterprise is built to turn decisions into action. Organizations fail at new technology because they invest in the first pillar and neglect the other two.

The first article resolved this into a single picture. Place an organization on two axes: the Arsenal on the horizontal axis, and the alignment of the Doctrine and the Formation on the vertical axis. Spending on capability moves you right, not up. Only work on the Doctrine and the Formation moves you up, toward Future-Proof, and most heavy AI spenders sit in the Automated but Brittle quadrant, having bought the tanks and kept the doctrine. This article is about the work of climbing out of that box.

One expectation to set before we begin. Future-proofing is a large and layered subject, and a treatment that claimed to be comprehensive would not survive the reader’s scrutiny. We focus instead on the moves that, in our research and experience, most decide whether an organization climbs or stalls.

II. The adaptations AI demands

The matrix tells you where you sit. This section is about the moves that carry you upward, toward Future-Proof. They are not Arsenal moves. Buying more capability slides you to the right, leaving you where you started on the vertical axis. The work that lifts an organization is work on the Doctrine and the Formation, and each adaptation below has an ancestor in 1940.

The newest evidence points the same way. In a July 2026 McKinsey Quarterly survey of 750 AI-using employees and leaders, 70% of respondents felt personally ready to work with AI, while only 27% of leaders said their organizations were ready to make the shifts required, and organizational readiness explained nearly twice as much of the difference between capturing value and not as personal readiness did. That gap has a consequence most leaders have not priced in: employees are not waiting. They are already using these tools in daily work, often on personal accounts, often without telling anyone. The choice in front of leadership is not whether AI enters the workflow, only whether it enters governed or ungoverned.

Doctrine: how leaders have to change

Shift 1: From chess master to gardener. This is how a leader actually runs Boyd’s loop at the edge. One of us made this case in the book Organizational Velocity through the words of an Army general who described how his own role had to change as he rose. The chess master reaches onto the board and moves every piece. The gardener sets the conditions, tends the soil, gives focus and direction, and then lets the leaders who own the ground run it. The general’s point was that only the person at the top can create the doctrine that pushes decision-making to the edge, the instinct the Germans institutionalized in 1940. What has changed is that the gardener’s stance is no longer merely good leadership. It is now a throughput requirement. As AI collapses the time it takes to execute, the leader who still inspects every task becomes the bottleneck the whole team waits on. The job moves from editor of the work to setter of direction.

Shift 2: Exploit and explore at once. The concept of organizational ambidexterity, the ability to explore new opportunities while exploiting existing operations, was introduced 50 years ago by Robert Duncan, but realizing the vision has remained elusive for most. Leaders have always been rewarded for running the core efficiently, the game of chess at which incumbents excel. The pressing demand is to keep playing that game while also fighting a faster and less predictable one, closer to mixed martial arts than to chess, and to keep the second from being starved by the first. Exploitation funds the quarter. Exploration is the only thing that funds the decade.

Shift 3: Manage the productivity illusion. Efficiency is not productivity. A tool that lets one team move faster can flood the teams downstream with more output than they can absorb, turning local speed into systemic drag. Worse, cheap generation invites pseudo-work, and confident machine output invites people to stop thinking, a pattern long documented as automation bias. Doctrine in this environment has to measure outcomes rather than activity and has to defend human judgment as deliberately as it once defended cost.

Shift 4: Treat smart failure as tuition. Can you identify the risk-takers in your organization by the arrows in their backs?  When pilots are judged by what was accomplished rather than what was learned, they are neither decisively killed nor decisively scaled, because both moves carry career risk. The result is the pilot purgatory and the why gap we described in Part One: projects idle with vague success criteria, and organizations record what was decided but almost never why. This makes every failure unusable as a lesson and every success unrepeatable as a practice.

Amy Edmondson, in her book, Right Kind of Wrong: The Science of Failing Well, draws the distinction that doctrine has to absorb. Intelligent failures, the kind that occur in new territory, at modest scale, in pursuit of a clear goal, are not evidence of bad management. They are the tuition an organization pays to learn faster than its competitors. Leaders who want exploration must price that tuition into the budget, say so out loud, and then keep their word the first time a well-designed experiment comes back empty. One punished smart failure teaches the whole organization to stop experimenting, and an organization that stops experimenting has quietly chosen the left side of the matrix.

These four shifts share a destination and a vulnerability. Each one changes how leaders decide, and not one of them survives contact with an organization built to resist it. A gardener’s doctrine inside a chess master’s structure produces frustration, not speed. Culture, in the end, is simply what the Formation makes normal. That is why the shifts above are only half the work, and why we turn now to the other half: rebuilding the Formation so the new doctrine has somewhere to live.

The gardener metaphor carries one further duty, and it may be the hardest: the gardener protects the new growth. Transformative technology is never neutral inside an organization. It shifts power, and the people thriving under the current system rarely attack it openly. They starve it, slow it, and fold it back into the routines it was meant to replace. In 1861, the Union’s chief of ordnance, Brigadier General James Ripley, rejected breech-loading and repeating rifles partly out of fear that soldiers would waste ammunition. When Wilder’s Lightning Brigade finally carried Spencer repeaters at Hoover’s Gap in 1863, the volume of fire was decisive. The pattern has not aged. In July 2026, Ukraine dismissed Mykhailo Fedorov, the defense minister behind its drone-centric transformation, after his push for rapid modernization collided with the established military leadership. Doctrine and formation do not defend themselves. Protecting the new, with patience in appropriate amounts, is a leadership act, not a sentiment.

Sometimes protection means separation, such as a unit with its own metrics, funding, and tolerance for failure, far enough from the core that the antibodies cannot reach. That structure produced the IBM PC in Boca Raton and Lockheed’s Skunk Works. But the bill comes due at reintegration. The PC business and GM’s Saturn both flourished apart and were smothered when folded back in, so the plan for bringing the work home deserves as much design as the decision to set it apart.

Formation: how the structure has to change

The structural adaptations follow the same logic. Top management identifies in clear and unambiguous terms the what: the outcomes to be achieved. The decision-makers, the people who are actually touching the problem, determine the How: the way those outcomes will be achieved.

Authority has to move to the nearest capable decision maker, whether human or machine, which is the organizational version of putting a radio in every tank. The discipline this requires is knowing what to control and what to release. Finances stay central because people at the edge see only their slice of the whole, but most operating decisions belong at the edge, exercised with freedom within clear boundaries. Work then must be carried out by mission-oriented pods, cross-functional teams that own an outcome end to end, because AI does not respect the org chart, and a fragmented response cannot keep pace with an integrated problem. A pod is small enough to name: an inbound-exception pod might hold a planner, a transportation analyst, a toolsmith, and the agent itself, own the exception rate as its one metric, and exist because three standing coordination meetings were killed to fund it. That is combined arms, not infantry support. And every intention here needs a mechanism, a written narrative, a weekly review, a metric with an owner’s name on it, because intent without a mechanism is just talk.

Doctrine: how leaders have to change

None of this is safe without a governance layer. Every automated decision needs a named owner, a control that defines when the decision is out of bounds, and an off switch that cleanly returns the work to people. Without all three, you do not have an automated decision. You have a science project running in production.  One caution on the off switch: it only works if the people still exist. An organization that automates the work and then cuts the manual capacity has an off switch wired to nothing and will discover it mid-outage.  Don’t forget that ownership scales to the top. Accountability for the AI-run operation sits with the operating leadership that owns the outcomes, not with a newly minted title at the edge of the org chart. And the edge can only be trusted with authority if signals, including bad news, travel without being filtered, and if people remain free to question the machine. We have argued in these pages that readiness turns on exactly this, what we call Accept readiness: whether the messengers of bad news are thanked or punished, and whether candor is safe. That openness is also the antidote to context debt, the accumulating risk that builds when teams can no longer explain, audit, or reproduce the rationale for a decision.

Automated is not autonomous

One distinction has to survive contact with vendors. An automated system runs a governed process tightly and repeatedly. An autonomous capability is many such processes orchestrated together, trading off service, inventory, and cost in real time. Amazon is the clearest illustration. A demand-forecasting system, a warehouse-robotics system, and a last-mile routing system each run independently and are governed and owned by their own organization; none is autonomous, but the network behaves autonomously only because the three are coordinated.

Unilever carries the logic to the operational edge, where agents, generative tools, and robotics now run a largely touchless logistics operation. The lesson is not the logos. It is that gains land first where data is abundant, decisions repeat, and handoffs are clear, and the durable advantage is the workflow, governance, and operating memory an organization accumulates. The model can be rented from anywhere. The system of work is the only part that is yours.

III. The skills layer

The tank story has a workforce ending we have not yet told. Germany’s advantage was built into its people. Its panzers carried radios, and its junior officers were trained to act on what they saw, so the army had developed the human capability its new way of fighting required. France had not. Its tanks were equal or better, many more heavily armored, yet most carried no radio, decisions waited for headquarters, and the army had spent a fraction of its budget on communications because the battle it planned to fight did not need them.

The reformers who saw the new fight coming—de Gaulle in France, Fuller and Liddell Hart in Britain—were largely ignored by their own establishments. The lesson is blunt. The new way of fighting was unavailable to any army that had not developed its people to wage it, however good its tanks. An organization that buys Agentic AI and never develops the leaders and workforce to question it, govern it, and redirect it is standing exactly where France stood in 1940, holding the better arsenal and unable to use it.

This human development capability is what this section is about. The future-proof organization is technology-enabled and human-led. The scarce resource is shifting away from hiring more data scientists and toward building AI-fluent operators who can question a recommendation, govern an automated decision, and redesign a workflow before handing it to a machine. The leading organizations are already creating exactly these hybrid roles, from human-AI collaboration managers to supply-flow analysts and engineers, and naming a new archetype, the versatilist, who reaches up and down the value chain to bridge the AI and the business. The fear that AI simply erases jobs is, so far, not what the evidence shows. Through the end of the decade, the projected job gains in asset-intensive industries outrun the losses. The work is changing faster than it is disappearing.

The tank offers one more instruction here, on specialization. France crewed its best tanks with a one-man turret: the commander directed the tank, watched the battlefield, aimed the gun, loaded it, and fired it. Germany split the same work three ways: commander, gunner, and loader, and each crewman, less important alone, was decisive as part of the team. When a task outgrows one person, the answer is to divide it deliberately, not to pile it onto the decision maker. The same division is emerging around AI: the decision maker who owns the outcome, the toolsmith who knows what the technology can and cannot do and keeps pace as it changes, and the data steward who ensures the question being asked is one the data can answer. Handing a manager an evolving AI toolset on top of an unchanged job is the one-man turret all over again.

The clearest example is a single role in transition, from data stitcher to trust builder. The old job was extraction and reconciliation: pulling data across systems, squaring the discrepancies, building the spreadsheet, and assembling the dashboard. It was perhaps 80% of the effort and almost invisible to leadership. The new job is verifying and defending what the machine recommends, applying judgment at the override point, calibrating when the machine can be trusted, and translating its output into a human context. That is where decision value is created once the routine stitching has been automated away.

The transition has a mechanism, not just a destination: pair the stitcher with the toolsmith. The people who spent years reconciling the data know better than anyone where it lies, and apprenticed to someone who knows what the machine can do, they become its natural verifiers.

Figure 1. A role in transition. As routine data stitching is automated, decision value shifts toward verifying, governing, and translating what the machine recommends.

The demand looks different at each level. The C-suite’s job is to set intent and guardrails, reallocate the portfolio, redesign incentives so the new behavior is rational, and protect exploration from starving. That is the gardener’s work, not the chess master’s. Mid-management is the translation layer, and the hardest gap to close. It calls for change leadership, cross-silo influence, and the discipline to redesign a process before automating it. The role is now under acute strain because AI ships finished work faster than a manager trained for a weekly cadence can review it. At the front line, the premium is on technological fluency, data analytics, and tolerance for ambiguity. Strikingly, most of what this transition demands is not technical at all. The scarce capabilities are people-and-change capabilities, which is the opposite of what most leaders assume when they first reach for technical hires. A scope note is owed here: this section addresses the decision layer, from planners to the C-suite. The physical workforce, in warehouses and behind the wheel, is being reshaped on a different trajectory, driven as much by robotics as by AI, and it deserves its own treatment.

Skills are built the way strength is. Start where people are, match the capability to the task in front of them, train progressively, and accept that what is not used is lost. And tie the investment to enterprise value, to cost, cash, growth, and risk, or it will not survive the next budget cycle. This is the through-line of our larger work on organizational velocity and data visibility. The binding constraint on AI is rarely the model. It is whether the people around it have been developed well enough to use it. Importantly, AI fluency is built at the keyboard, not in the briefing. A leader who has never personally put these tools to work cannot judge what to trust them with.

IV. How to act: Can I, should I, do I want to?

Knowing where you sit is not the same as knowing what to do about it. The move from one box to another runs through three questions, asked in order. They are not a new framework. They are how an organization climbs the matrix, and they extend the readiness dimensions we proposed in these pages: whether an organization can Observe change, Accept it, and Act on it. Can I tests the capacity to act. Should I asks whether you have observed the landscape honestly. Do I want to asks whether the organization will truly accept the change.

Can I is a question about capacity, and beneath the infrastructure and the budget it is really a question about data. Test it against what we call the data readiness trinity. Quality: is the data clean and trustworthy enough to act on? Context: do functions and partners share definitions, so the data means the same thing to everyone? Access: does the right data reach the right people at the right time, through governance rather than around it? The anti-pattern is data wishcasting, the assumption that the data will materialize once the project begins. If you cannot answer basic supply chain questions with confidence today, you are not ready for advanced capability, however good the tool.

Should I is a question about strategic posture. Not every organization must lead every change, and letting others go first is a legitimate and often winning move. Apple was not first to digitize music, and Netflix moved deliberately from discs to streaming to content. The test is twofold: whether the shift merely complements your business or stands to transform it, and whether you own the layers needed to solve your hardest problem or are renting them. If you are renting the layer that will decide your future, that is the signal to stop waiting and move. And first-mover advantage, where it exists, belongs less to the first company to own a technology than to the first to recognize how it converts into value for its customers, and to act on that recognition ahead of the field. There is a pattern behind that test, one Clayton Christensen described. When a technology frontier shifts far enough, the modular, off-the-shelf pieces that served the old performance level stop being good enough, and integration beats modularity again until performance once more overshoots what customers actually need. We are in one of those not-good-enough-again moments, which is why the organizations pulling ahead are reintegrating and owning the layers that decide their hardest problems rather than renting them. This is a lens, not a mandate to vertically integrate everything.

It also resets the point at which the advantage is won. Technology competes on three rungs: doing the familiar job better, faster, and cheaper; serving a need that everyone currently meets poorly; or serving a need no one is meeting at all. Most organizations assume they are playing on the first rung. Should I, answered honestly, includes asking which rung you are actually playing on.

But fast-following carries a catch that the strategy literature usually misses. You can fast-follow on the Arsenal. You cannot fast-follow on the Doctrine and the Formation. And there is a third thing you cannot fast-follow at all: the learned ability to exploit the openings that Doctrine and Formation create. That capacity accumulates; it cannot be purchased. Every experiment in the portfolio is a repetition, every post-mortem converts a miss into a lesson, and every recorded why turns experience into the operating memory a competitor cannot acquire on the timeline of a procurement cycle. It may prove the most durable advantage of the three.

Do I want to is the question most organizations skip, and it has two parts. The first is appetite, the willingness to accept short-term risk and failure for long-term gain. The second is incentives, which are part of the Formation and quietly govern everything else. If the reward system still pays for yesterday, people will keep working yesterday, whatever the strategy deck says. Performance measurement is the silent killer of transformation because it works without anyone deciding anything. One of us watched this for years at UPS, where return on invested capital governed executive rewards. The metric was meant to measure investment skill, but the surest way to improve it was to invest as little as possible, and with executives holding a role for only a few years, the certain path beat the ambitious one every time. The discipline to hold here is a paradox: manage for stability and plan for change at the same time. And planning for change means changing the management style and the structure, not only the technology.

The apparatus that makes this practical is to treat AI as a portfolio of decisions rather than a stack of products, and to govern, stage-gate, and defend it to the CFO as a portfolio. Define the minimal viable data for each use case, the smallest reliable data set that lets it ship and improve, rather than boiling the ocean, because AI in production cleans data faster than committees ever will, provided someone owns the correction loop. Left ungoverned, it automates the errors instead. Move each use case through gates, from explore to prototype to shadow mode to production, and treat every gate as a financial decision with a baseline, a shadow-versus-control comparison, an accuracy measure, a read on service and cost, and pre-agreed kill criteria. As one executive told us, do not bring a fifty-million-dollar proposal before you have a ten-thousand-dollar truth. And never forget that if you automate a broken process, all you get is broken faster.

V. The diagnostic: Questions to ask before the next check

The argument of this two-part article reduces to a small set of questions a leadership team can ask itself. They are deliberately uncomfortable, because the comfortable version of each one is exactly how organizations end up Automated but Brittle. Answer them honestly, with the people one level down in the room, before authorizing the next round of spending.

Honesty: where we actually stand, and why we are buying

• Where are we actually on the path from manual work to autonomy, and would the people one level down from us give the same answer?
• Are we buying this technology to amplify basics we have mastered, or to compensate for basics we have not?
• Are we buying this technology simply because everyone else is and we are afraid of missing out (FOMO strikes again)?
• Can we answer our hardest supply chain questions with confidence today, before layering advanced capability on top?

Doctrine: how we decide

• Do our leaders work across functions and influence decisions beyond their direct authority, or do they execute within fixed lanes?
• When the machine runs faster than our approval chain, who is empowered to act at the edge, and within what boundaries?
• Are the messengers of bad news thanked or punished, and is it safe to question the machine?
• Do we differentiate between intelligent failures that help the organization learn faster than competitors and failures that either should have been known in advance or do not create any new learnings?
• Are we essentially adding more work to the manager or have we redesigned the process to allow the manager to better do their jobs through the use of the new tools?
• Do we allow experimentation and do we learn from such experimentation?

Formation: how we are built to act

• Does authority sit with the nearest capable decision maker, human or machine, or does it wait for headquarters?
• Do our incentives still reward yesterday’s behavior, and if so, why would anyone change?
• Does every automated decision have a named owner, a defined out-of-bounds condition, and a kill switch?
• Do we recognize the importance of and need for task specialization?
• After a failure, do we sit down and ask what went right, what went wrong, and what was missing? Have we developed an environment of trust and mutual respect?
• Have we built a culture of complacency, or one that keeps pushing to do better? Do we understand the concerns and fears of our personnel, and have we developed plans or approaches to deal with these issues?
• Does this work need a protected structure with its own metrics, and what is our plan for bringing it back home?

The why and the will

• Do we record why decisions are made and which alternatives we rejected, or only what we decided?
• Are we willing to unlearn the practices that made us successful, even the ones we are proudest of?
• Do we recognize that managers who were good at one set of skills or outcomes may not be able to transition to the new environment and new outcomes, and do we have procedures in place to work with these managers?
• Have we proven we can generate real value in a small pilot before committing the organization to an expensive path?

VI. The choice

In the end, an organization that refuses to change is making a quiet demand of the market: that customers change to its way of doing things rather than the other way around. A strong brand can enforce that demand for a while, on a slice of the market that keeps getting smaller.

The organizations pulling ahead are doing the opposite. They are doing the work technology cannot do for them, changing how decisions get made and who gets to make them, and resetting the incentives underneath.

That work has a shape, and it is not the one most change programs assume. The hard part is not learning the new. It is unlearning the old, the practices and assumptions that earned their place and now quietly block what comes next. Learning is additive and comfortable. Unlearning is subtractive and threatening, because it asks an organization to let go of the very things it is proudest of. Only after the unlearning can the relearning take hold. That sequence, learn, unlearn, relearn, is the real work of future-proofing, and unlearning is the hinge the whole thing turns on.

Return, one last time, to France in 1940. The French were not surprised by the tank. They built more of them, and better ones. They were not short of warning either. De Gaulle had published the case for concentrated armored formations years before the Germans proved it. The reformers were in the room and the tanks were in the field. What was missing was the willingness to unlearn a way of fighting that had won the last war. The bill for that refusal came due in six weeks.

Your bill will not arrive that fast, and that is the danger. It will arrive quarter by quarter, as a rival’s forecasts get a little sharper, its exceptions resolve a little faster, and its people spend a little more time on judgment and a little less on stitching data together. None of those gaps will look decisive on its own. Then one day the gap is no longer a gap. It is a different league, and the years it would take to rebuild your Doctrine and your Formation are exactly the years you no longer have.

Here is the good news. Everything your competitors can buy, you can buy. The Arsenal is for sale to everyone, which means it is a lasting advantage to no one. The Doctrine and the Formation are not for sale at all. They are built, deliberately and over years, by leadership teams willing to question what made them successful. That is the one part of this transformation that cannot be copied, cannot be rented, and cannot be fast-followed. And it sits entirely within your control.

The question was never whether to adopt AI. It is whether you will do the work that lets it matter. The diagnostic above is where that work begins: the questions asked honestly, with the people one level down in the room, before the next check is signed. France had six weeks. You have longer. Not much longer, but enough, if you start now.

About the authors

Steven A. Melnyk, Ph.D., is an emeritum professor of Supply Chain Management at Michigan State University. He has co-authored 25 books, more than 100 refereed journal articles, and more than 300 practitioner articles. He is a past member of the APICS Board of Directors. Since retiring in 2024, Dr. Melnyk is now part of the Executive Development Program at Michigan State University and can be reached at Melnyk@broad.msu.edu

Alan Amling, Ph.D., is an assistant professor of practice in Supply Chain Management at the University of Tennessee’s Haslam College of Business and CEO of Thrive and Advance, LLC. This article is adapted from “Organizational Velocity: Turbocharge Your Business to Stay Ahead of the Curve.” He can be reached at aamling@utk.edu.

Executive takeaways

Selecting a warehouse management system (WMS) begins with understanding the business requirements, evaluating available products, and making a build-versus-buy decision. That evaluation should be rigorous because once implementation begins, the real test is how well the selected product fits the operation.

Even after careful selection, however, implementation teams will find gaps. A requirement may not be supported by the WMS out of the box, or a legacy operating process may not align with the way the standard product works.

When that happens, three choices emerge. The business can change its operational process to align with the standard WMS—a process transformation. It can extend the product using mechanisms provided by the vendor. Or it can customize the product itself.

The difficult questions are not simply whether customization is good or bad. Does the existing process truly need to be preserved? Can the requirement be met through an extension? When is customization justified? And if customization becomes extensive, should the organization revisit whether it selected the right product—or whether it should have built instead?

The chicken pasta test

Consider a simple analogy. You order chicken pasta from a restaurant because you want to save the time and effort of preparing it at home. After receiving it, however, you make a separate curry, mix it into the pasta, and perform several other steps to turn the dish into what you actually wanted. At some point, it is reasonable to ask: Why did you buy the pasta in the first place? If you were going to spend that much effort changing it, could you have prepared the entire dish yourself?

The same logic applies to WMS selection. A purchased WMS does not have to meet every requirement without any change, but an organization should be able to use most of the product as designed. If the implementation requires the team to repeatedly rebuild standard behavior, the economics and logic of buying the product begin to weaken.

Detailed requirements and product evaluation reduce this risk. Critical use cases should be understood at the operational level and evaluated during product selection so that major gaps are identified as early as possible.

There is no universal percentage that tells an organization how much customization is acceptable. The answer varies by business and by requirement. What matters is having a disciplined way to decide among process transformation, extension, and customization—and recognizing when the level of change is becoming so extensive that the original product decision should be questioned.

Heavy customization also creates technical debt. Custom code and changes become the customer’s responsibility to maintain. If they interfere with standard functionality or fail after an upgrade, the vendor may not be able to support them in the same way it supports the standard product. At some point, extensive customization can defeat one of the primary reasons for buying a packaged WMS in the first place.

Three ways to address a WMS gap

When a requirement is not supported by the standard WMS, the response generally falls into one of three categories:

1. Transform the business process so that it works with the standard WMS functionality.
2. Extend the WMS using vendor-provided extension points or hooks. This is typically a more moderate-risk approach.
3. Customize the core product. This carries substantially more risk and should be reserved for cases where it is genuinely necessary.

Process transformation is usually the best path—but not a rule

Process transformation is generally the preferred path because it preserves the standard product. But can every business process be transformed to fit the WMS? No.

Some organizations have genuinely unique requirements that the standard product does not support. In other situations, process transformation is possible but may take significant time. The implementation team then faces a trade-off: Should the WMS program wait for the business process to change, or should the team extend or customize the product now, launch the WMS, and transform the process later?

Two broad situations can emerge:

1. Process transformation will take a long time, while the ROI from implementing the WMS now is high. In that case, an extension or even a carefully justified customization may be worth the risk.
2. The required customization would fundamentally alter the product and create unacceptable operational or technical risk. In that case, the customization should not be pursued; the organization should transform the process instead. A critical mismatch of this scale should also prompt the team to revisit the assumptions made during product selection.

Extension and customization are not the same

The distinction between an extension and a customization is important within the scope of how it’s in this article.

A customization changes how the available product functionality itself works. For example, assume serialization is enabled for a particular item in the WMS, but when that item is picked and shipped under a particular condition, the business wants serialization disabled. If the WMS does not support that behavior and the customer changes the product code to make it happen, the customer has customized the product.

This is risky because standard functionality can have many dependencies that the customer may not fully understand. Even if testing is successful, the change can create problems elsewhere or break later. The vendor may also be unable to support the modified behavior because the customer has effectively created its own version of that functionality. During upgrades and patching cycles, the organization may also be responsible for preserving or reworking those customizations.

An extension is different. WMS vendors recognize that a generalized product cannot natively meet every requirement of every customer or industry. They therefore provide extension points or hooks that allow customers to add their own business rules without changing the core product.

Consider label printing during a pack transaction. The standard WMS may print a label for the parent LPN or handling unit. One customer may also need labels for the child LPNs. Another may need parent and child labels only for particular customers. If the vendor provides an extension point that allows each customer to implement these rules while preserving the standard pack functionality, the product can be personalized without changing its core.

This is an important consideration during product selection. Do not evaluate only the strength of the standard functionality. Evaluate how well the product can be extended without core customization. A strong extension framework can allow an organization to preserve the integrity of the standard product while still accommodating unique business requirements.

Get to this level of detail before product selection

If a requirement can be met through a vendor-supported extension, there may be no need to choose between core customization and process transformation. Ideally, these possibilities are understood during product selection, although real implementations can still uncover operational complexity that was not fully visible during an RFP.

A disciplined sequence is:

1. Document detailed business requirements down to the ground-level operation, including the different scenarios the WMS must support.
2. Write detailed test scenarios so that anyone reviewing them can clearly understand what is expected from the WMS at the lowest practical requirement level.
3. Share those requirements and scenarios with the WMS vendors under consideration.
4. Analyze how each product meets the requirements through standard functionality, supported extensions, or customization.
5. Use that analysis to make the build-versus-buy decision and, if buying, select the best-fit product.

Example 1: When process transformation is better than customization

Consider a distribution warehouse that receives material from suppliers, stores it, and ships it either to manufacturing plants or directly to end customers as spare parts. The same item can therefore be used as a manufacturing component or shipped as a spare.

The organization multi-sources the item from several suppliers to reduce single-sourcing risk. For historical reasons, some suppliers transmit serial numbers and others do not. When the item is shipped to a manufacturing plant in full pallets, the warehouse does not need to scan the serial numbers. When the same item is shipped as a loose spare, however, serial numbers must be captured.

The existing homegrown WMS was built around this legacy process. Suppliers 1, 2, and 3 do not provide serial numbers, while suppliers 4, 5, and 6 do. The WMS therefore maintains serialization at an item-and-supplier level: inbound receipts from suppliers 1, 2, and 3 are allowed without serials, while serials are mandatory for suppliers 4, 5, and 6. When a spare is shipped and the serial was not captured at inbound, the warehouse captures it during outbound processing.

This is a highly specific serialization profile, and a standard WMS may not support it. Assume the organization has nevertheless chosen to buy a WMS because the product meets many other requirements. For this serialization gap, the organization has two choices: customize the WMS to reproduce the existing process, or examine whether the process itself should change.

The second option exposes an inefficiency. Why are suppliers 1, 2, and 3 not sending serial numbers when suppliers 4, 5, and 6 already do for the same item? If the organization changes the process and requires all suppliers to transmit serial numbers, the complex WMS requirement largely disappears.

The item can be serialized at the enterprise level rather than at the item-and-supplier level, and the warehouse no longer needs a special outbound serial-capture process based on destination because the serial number is already available when the product is received.

This is a strong example of process transformation accomplishing two objectives at once: it improves warehouse efficiency by eliminating the labor required to capture missing serial numbers, and it aligns the operation with the standard WMS product.

But process transformation also has a cost

Every unsupported requirement should be examined through this lens, but process transformation is not automatically the answer. Some requirements are legitimate and cannot reasonably be transformed. In other cases, transformation is correct in principle but impractical within the WMS implementation timeline.

Return to the serialization example. Changing three suppliers may be manageable. But what if 300 suppliers must change their processes and begin transmitting serial numbers? Implementing one WMS would now depend on changes across 300 external organizations.

The organization must compare at least two things:

1. How long will it take to change all 300 supplier processes, and what value will be lost if the WMS implementation is postponed during that period?
2. What will it cost to customize the WMS now and maintain that customization until the supplier transformation is complete?

Suppose the customization costs $100 to implement and another $100 per year to maintain for two years, for a total cost of $300. If waiting two years for the supplier transformation would cause the organization to forgo $3,000 in value because it must continue operating with poor data quality and suboptimal supply-chain performance, then spending the $300 may be the better decision.

The specific numbers are only illustrative. The point is that the decision should be made by comparing the cost and risk of customization with the value lost by delaying the WMS and waiting for process transformation.

Example 2: When a legitimate business requirement may require customization

Now consider a build-to-order manufacturer in which individual products can be different from one another. Windows and doors are a classic example. Assume the company is fulfilling a large order for a luxury shopping complex containing 1,000 windows and doors, many of them unique, all shipping to the same destination.

Each window or door may be represented as its own LPN or handling unit, while several individual units are wrapped together on a pallet for shipment. This creates a nested handling-unit structure in which multiple child LPNs sit within a parent LPN.

If nested handling-unit picking is a legitimate and essential part of the business process and the standard WMS does not support it, the organization may have to customize the WMS.

If this is a make-or-break capability, it should ideally have been discussed with WMS vendors during product evaluation and considered in the build-versus-buy decision. If the organization still chooses to buy and customize, there should be a clear reason—such as the fact that the product satisfies many other important requirements through standard functionality.

Where to draw the line

The decision starts with understanding what the business actually needs the system to do. For each gap, ask: Does the current process deserve to be preserved? Can it be transformed? Can the requirement be handled through a vendor-supported extension? If core customization is truly necessary, are its cost, maintenance burden, operational risk, and business value justified?

The goal is not to eliminate customization at all costs. It is to avoid customizing a WMS merely to preserve a process because "this is how we have always done it." At the same time, organizations should not force process transformation when a requirement is legitimate or when waiting for transformation would destroy more value than a controlled customization would cost.

That is ultimately where the line should be drawn: do not customize software to preserve a process until you have established that the process deserves to be preserved. When transformation, extension, and customization are evaluated deliberately, organizations are far more likely to preserve the value of a packaged WMS while still meeting the needs of the business.

Although the examples here focus on WMS, the same principles can be applied to other enterprise software implementations.

About the author:

Sreekumar Somasundaram is a senior supply chain technical program manager at AWS, specializing in warehouse and transportation management systems implementation. With over 18 years of experience, he has been involved in developing and implementing logistics technology solutions across complex business environments. A mechanical engineering graduate with multiple professional certifications including Project Management Professional (PMP), Java and Oracle Programming, Transportation Management, and Supply Chain certifications from MITx, he brings a unique combination of technical expertise and business acumen to his role. His extensive background encompasses logistics solutions architecture, business analysis, and software development for warehouse and transportation systems. Throughout his career, he has successfully led WMS and TMS implementations for multiple Fortune 500 companies.

Today is Monday, Sept. 28

1. Record tanker costs push Trafigura toward vessel ownership

Trafigura is moving from chartering to owning tanker capacity through a new 14-vessel business as crude-shipping rates reach record levels, illustrating how scarce transportation assets can reshape sourcing and capital-allocation decisions.

FreightWaves  |   Read the original article

2. Bangkok floods slow container and intermodal flows

Severe Bangkok flooding is disrupting road access, container movements and rail links to Laem Chabang Port, with logistics providers warning of longer pickup, delivery, gate-in and equipment-return times.

Kuehne+Nagel myKN  |   Read the original article

3. Mumbai congestion forces Lufthansa freighter pause

Lufthansa Cargo suspended dedicated freighter service to Mumbai through September as construction, seasonal volumes and handling backlogs produced truck waits of up to three days and delayed cargo acceptance and customs processing.

FreightWaves  |  Read the original article

4. Arctic cargo traffic tests a new strategic corridor

Arctic cargo-vessel movements reached a four-year high in August as Russia expands energy exports and Asian operators gain experience, although seasonal limits, sanctions, insurance costs and environmental concerns constrain wider adoption.

Kuehne+Nagel myKN / Lloyd's List  |  Read the original article

5. Maritime networks move higher on the cyber-risk list

An EU threat assessment found transportation drew 11.7% of identified cyberattacks last year, with maritime networks exposed by legacy systems, operational technology and removable media as China-linked campaigns targeted shipping organizations.

Kuehne+Nagel myKN / Lloyd's List  |   Read the original article

6. Food manufacturers tighten procurement and planning

Food manufacturers outlined cost and resilience moves spanning procurement consolidation, freight optimization, AI-supported forecasting, SKU reduction and sourcing changes as inflation and uneven transportation rates continue pressuring margins.

Manufacturing Dive  |   Read the original article

7. Cargo thieves move upstream into freight systems

The FBI’s Memphis Cargo Theft Task Force says criminals increasingly compromise accounts and freight platforms before trucks arrive, using shipment data and fraudulent identities to redirect high-value loads.

FreightWaves  |   Read the original article

8. Modular gas monitor targets changing industrial hazards

Honeywell introduced a cartridge-based portable gas-detection platform that allows industrial teams to change sensing configurations in the field for applications including VOC screening, emergency response and leak detection.

Work Safety 24/7   |   Read the original article

9. Gas-remediation system pairs detection with removal

NanoSieve says its integrated system can detect and remove selected flammable, toxic and greenhouse gases in real time, targeting industrial safety applications involving methane, carbon monoxide, hydrogen and battery off-gas.

Work Safety 24/7   |   Read the original article

Demand forecasting often relies on planner judgment, manual intervention and fragmented data. Bayer and ZS are reimagining that model through data science-led, exception-based forecasting that expands no-touch planning and focuses planner expertise on higher-value decisions.

Supply chain and planning leaders will hear how Bayer is scaling machine learning across stable products, new product introductions (NPI), loss of exclusivity (LOE) and tenders, and what it takes to embed data-driven forecasting in everyday planning.

In this webinar, you'll learn:

• How to shift from planner-led forecasting toward a data science-led, exception-based model
• How Bayer is pursuing no-touch forecasting across as much of its portfolio as possible
• How machine learning can be applied across distinct forecasting challenges, including stable, NPI, LOE and tender portfolios
• How to redefine the planner’s role so human expertise is focused where it creates the greatest value
• Why scaling AI and machine learning requires transformation across people, process, data and operating model, not just technology

FEATURING:
Caglar Ozdag, Principal, Supply Chain Planning, ZS and Andreas Schriber, Director, Customer Planning Strategy and Capabilities, Bayer

Executive takeaways

Talking politics at work has always been thorny. But today, it’s increasingly impossible to avoid: From tariffs to AI regulation, immigration policy, climate legislation, labor laws, or sanctions, more and more seemingly apolitical business decisions have become politically charged—with major consequences for how well those decisions get made.  What does it take to lead effectively in this environment?

Most executives already know that politics can divide teams. What they don’t always realize is that political identity can quietly reshape business judgment, even when a decision isn’t explicitly political. Employees may believe they are evaluating the merits of a proposal, when they are actually being heavily influenced by their evaluation of whether the person proposing it shares their worldview. In other words, left unchecked, political bias means that business decisions risk becoming identity decisions.

To understand the impact of political identity on seemingly apolitical decisions, we conducted an experiment with 216 supply chain managers. All the participants in our study evaluated exactly the same business proposal: a manager recommending changing suppliers in response to tariffs, with an independent estimate showing modest cost savings. Some participants read a politically neutral explanation of the proposal, while others read the same proposal accompanied by either a pro-Trump or an anti-Trump comment from the manager. We then asked the participants to rate their level of support for both the manager and the proposal.

The results surprised us: Employees consistently viewed managers who expressed political opinions less favorably, even when the employee shared their political views. In addition, when managers stayed neutral and avoided political statements, employees still formed assumptions about the managers’ political identity—and when they perceived ideological differences, they were less likely to support not only the manager, but also the business decision itself.

These findings demonstrate that once a topic becomes politically charged, employees don’t just ask whether something is a good business idea. They start to wonder about a different set of questions: Does this colleague see the world the way I do? Can I trust their judgment? Are they making this recommendation because it is good for the business or because it reflects their politics?

These questions are rarely explicit. But they can dramatically transform how people interpret the same facts. It’s not just that people disagree about policy. It’s that political identity can (unintentionally) become a shortcut for evaluating people and their ideas.

So, what does it take to keep identity judgements from unduly influencing business decisions? To explore how executives are navigating political polarization at work today, we conducted a series of follow-up interviews with leaders in aerospace, energy, logistics, pharmaceuticals, and consumer goods. Across these interviews, the executives we talked to consistently described four strategies that helped them lead through these challenges:

1. Accept that politics are part of business

Some leaders try to keep politics out of the workplace altogether. That instinct is understandable, but increasingly unrealistic. Organizations cannot ignore tariffs, AI regulation, immigration, climate policy, labor laws, or geopolitical conflict, because these developments directly affect customers, costs, supply chains, hiring, investment, and strategy. Ignoring them simply leaves managers unprepared.

Given this reality—and our key finding, that even seemingly neutral topics can become politically charged if the person talking about them seems to have a different political identity than we do—avoiding politics often just isn’t an option. As one executive from a consumer goods company told us, “Should politics be something you simply don't discuss in a company? My answer is no. Politics is everywhere. Every strategy process begins with a PESTEL analysis, and it’s no coincidence that the first letter stands for the political environment. You can’t be politically agnostic.”

Similarly, an executive at a pharmaceutical company described how “there is no sweeping [politics] under the rug. Think about vaccination policies during Covid-19 or today’s debates around AI regulation. You simply can’t avoid these issues.” After all, when political issues impact business decisions, “just don’t talk about it” isn’t a feasible strategy. As a president of a logistics company succinctly put it, “We have to figure out what to do about it. We run a business.”

2. Separate political developments from partisan identity

One of the most surprising findings from our study was that managers gained nothing by expressing their political opinions: Even when an employee agreed with their manager’s opinion, explicitly political statements reduced their support for the manager. In other words, even if you think you’re politically aligned with your team, it’s best to focus on the impact of political developments on the business, rather than on your partisan identity.

One supply chain executive captured this distinction in four words: “Politics matters. Politicizing doesn’t.” An energy executive described how this principle guides his communication: “When we publish information or present an analysis, we work hard to keep it to the facts. No added commentary. No ‘I wish it were this way’ or ‘I wish it were that way.’”

A senior sourcing manager shared a similar perspective, explaining, “There is a distinction between managing political policy and personal opinions,” while a logistics executive reflected that “It’s not about what we think Trump is doing. We have to figure out how to make the business successful.”

Employees need to understand how political developments influence business decisions. They rarely need to know which politicians or parties their manager supports. While we shouldn’t try to hide our political identities entirely, keeping the focus on business impact can help organizations stay on track and reduce the potential for partisan judgements to bias decisions.

3. Focus on the facts

Effective leadership isn’t about trying to eliminate disagreement. Rather than attempting to suppress political conversations, the executives we interviewed channeled these discussions toward their organizations’ objectives.

One aerospace executive put it simply: “Discussions around political issues are unavoidable. The challenge is not to prevent them, but to channel them productively.” He continued, “As a leader, I need to bring the conversation back to what the objective implications are for our business. It can be hard because I have my own opinions. But from a professional leadership perspective, I have to channel the conversation into objective decision-making that creates value for the business.”

This doesn’t mean banning political conversations at work. But it does mean guiding them toward productive, business-relevant topics. For example, one manager shared that it’s important to “acknowledge people’s opinions, but remind them that team meetings are not the place,” while another observed, “It’s about controlling the temperature of the room and keeping the conversation focused on the facts.” By redirecting attention from political identity back to business consequences, leaders can recognize political realities without allowing them to derail workplace decisions.

4. Build political resilience

Perhaps the most uncomfortable implication of our findings is that managers cannot eliminate political bias altogether. Even if you avoid making explicitly political statements, seemingly neutral topics can easily become politically charged, and you may still be perceived as politically different from members of your team when you talk about them. Moreover, people’s assumptions about your political identity may influence their evaluations of you and your work (and your own assumptions may influence your judgements, too).

As such, leaders need to take steps to mitigate these risks, but they also need to cultivate their political resilience to persevere in the face of them. One consumer goods executive summarized it this way: “Managers need to understand that discussions around political topics can affect both their personal reputation and support for the decisions they are trying to implement. Ultimately, managers still have to do what is best for the company. Part of leadership is accepting that not everyone will agree with you. You need resilience. Build your immune system.”

Another manager from the energy industry described the mindset that navigating these situations required: “You have to ask yourself, ‘Why am I here? Is it worth the argument?’ You’re not going to win somebody over if they’re that different from you. As long as this person is doing their job and our goals are aligned from a business perspective, then we have to continue.”

Political resilience is about recognizing that disagreement is inevitable, refusing to take it personally, and continuing to lead with professionalism and discipline. After all, our research demonstrates that effective leadership is not about avoiding political issues. It’s about acknowledging political realities and keeping them from interfering with good decision-making.

About the authors

Dr. Craig Carter is the John G. and Barbara A. Bebbling Professor of Supply Chain Management in the W.P. Carey School of Business at Arizona State University.  He is also an Advising Editor and Editor Emeritus of the Journal of Supply Chain Management and Fearon Fellow at CAPS Research.  Professor Carter has worked in the areas of transportation and logistics with Ryder Systems, Hechinger Company, and the U.S. Department of Transportation. 
 

Lutz Kaufmann is professor of business negotiations and procurement at WHU – Otto Beisheim School of Management in Vallendar, Germany. His research and teaching focus on procurement and negotiation strategies. He can be reached at lutz.kaufmann@whu.edu.

Scott DuHadway, Ph.D. is an Associate Professor of Supply Chain Management at Portland State University. Dr. DuHadway received a Ph.D. degree in Operations and Sourcing from Michigan State University in 2016, and a B.S. degree in both Operations Management and Business Administration from Utah State University in 2012. His research focuses on the interface between risk, technology, and organizational behavior with an emphasis on supplier relationship management. In particular, he studies what motivates firms to engage in intentionally opportunistic or deceptive behaviors.

Today is Friday, Sept. 25

1. U.S.-China trade pause gains two more months

The United States and China will extend their trade truce through Jan. 10, preserving reduced tariffs and paused trade actions while negotiations continue, although the White House had not formally documented the extension.

Supply Chain Dive  |  Read the original article

2. Planning autonomy remains a distant goal

Gartner expects only 5% of organizations pursuing planning automation to make at least 10% of decisions autonomously by 2030, citing continued gaps in data, technology, decision ownership and workforce readiness.

Gartner  |  Read the original article

3. Tariff refunds flow into cash and capital projects

An Atlanta Fed survey found most recipients are retaining at least some IEEPA tariff refunds as cash, while 52.5% are directing returned funds toward research, development or capital projects.

Supply Chain Dive  |   Read the original article

4. Amazon gives sellers AI help with inventory decisions

Amazon is adding AI agents for inbound planning and aged inventory to Seller Assistant, alongside consolidated shipment visibility and compliance tools intended to simplify sellers’ international inventory and fulfillment decisions.

Supply Chain Dive  |  Read the original article

5. Canada proposes new rules for supply-critical labor disputes

Canada’s proposed Bill C-39 would create a more structured process for government intervention in disruptive federal-sector labor disputes, drawing support from freight interests and strong objections from organized labor.

The Canadian Press via EverythingGP  |   Read the original article

6. EU frames a research agenda for stronger food chains

The European Commission’s AgRI 2040 conference is setting research priorities for more competitive, resilient and sustainable food systems, including stronger value chains, digitalization, artificial intelligence and robotics.

European Commission  |   Read the original article

7. Industrial inspection robots gain secure door credentials

ANYbotics, dormakaba and LEGIC developed secure door access for ANYmal inspection robots, enabling authenticated movement through restricted industrial areas while preserving site permissions, safety controls and auditable entry records.

Robotics 24/7   |   Read the original article

8. NVIDIA integration targets contact-sensitive robot assembly

Flexiv integrated its force-controlled Rizon 4 robot with NVIDIA Isaac, giving developers a simulation-to-production workflow for contact-sensitive insertion tasks where fixed trajectories can struggle with small positional variations.

Robotics 24/7  |    Read the original article

Executive takeaways

Most supply chain resilience work in high-tech manufacturing happens at the Tier 1 level. That’s understandable—Tier 1 relationships are the ones with names attached, contracts negotiated, and quarterly business reviews on the calendar. If you’re a CSCO or head of procurement, it’s also where you have the most leverage and the clearest visibility, so it’s natural that’s where the resilience budget goes.

It’s also, in my experience, not where most of the damage actually happens.

The disruptions that blow up a production schedule tend to originate two or three layers deeper, in the sub-tier suppliers providing memory, storage, and semiconductor allocation that a lot of Tier 1 partners are themselves dependent on and don’t fully control. DRAM and NAND lead times can move from weeks to months with little warning. Controller silicon allocation gets reprioritized industry-wide when a large customer places an outsized order. None of that shows up in a Tier 1 scorecard until it’s already a problem, because by the time it reaches your Tier 1 partner, it’s already someone else’s crisis that’s about to become yours.

I’ve sat with procurement teams who had genuinely sophisticated Tier 1 risk management—dual-sourced assembly, geographic diversification, solid contractual protections—and watched all of it get overrun by a NAND allocation cycle nobody on the team saw coming because nobody was looking that far down the BOM. A fully optimized Tier 1 strategy doesn’t protect you from a problem that starts three tiers below where your visibility ends.

Start by mapping where your BOM is actually vulnerable

The first step isn’t buying more inventory. It’s figuring out where you need to. Most complex bills of materials have a small number of components driving most of the risk—usually memory, storage controllers, or specialty semiconductors with concentrated supplier bases and long, allocation-driven lead times. Everything else on the BOM is comparatively replaceable or multi-sourced already.

The audit that matters here isn’t a generic supplier risk review. It’s component-specific: for each critical part, who actually manufactures it, how concentrated is that manufacturing base, what’s the current allocation environment look like, and what’s your actual exposure if lead times double. I’ve seen procurement teams get this wrong in both directions—either treating the whole BOM as equally risky, which makes the problem unmanageable, or assuming risk is evenly distributed across their Tier 1 base, which misses where it’s actually concentrated. The teams that handle this well narrow down to a manageable list of 5, 10, 15 components that deserve real attention, and treat everything else as lower priority.

Don’t choose between allocation agreements and buffer stock—use both, deliberately

Once you know what’s actually vulnerable, the next question is how to secure it. The instinct is often to pick one lane: either lock in direct allocation agreements with manufacturers, or build a buffer of inventory to ride out disruption. In practice, the more resilient approach usually blends both, and the mix should differ by component.

For components with a small number of qualified suppliers and long design-in cycles, direct allocation agreements are worth the negotiating effort—they buy you priority when a shortage hits, even if they don’t fully protect you from price movement. For components where you have more sourcing flexibility or where allocation agreements aren’t practically available at your volume, a buffer strategy does more work. The mistake is applying one approach uniformly across the BOM instead of matching the strategy to how each component’s supply base actually behaves.

Build the buffer without letting it wreck your balance sheet

This is where a lot of otherwise sound strategies stall, and it’s the piece I see under-discussed relative to how much it actually determines whether a resilience plan gets executed. Building a meaningful buffer of memory, storage, or allocation-constrained silicon means tying up real capital in components sitting in inventory, and depending on how a company is structured, that inventory shows up on the balance sheet in a way finance teams notice immediately—working capital, inventory turns, the ratios a board or lender is watching. A procurement team can have the right component-level risk assessment and still get the buffer plan scaled back or rejected because nobody worked out how to carry it financially.

This is a tension I deal with constantly at Wintec, where we manufacture key hardware components and help clients structure the inventory and financing side of decisions like this. The practical fix isn’t to avoid holding buffer inventory—for allocation-constrained components, that’s often not optional. It’s to structure how that inventory is held so it doesn’t force a tradeoff between component resilience and financial discipline. Companies that get this right treat the financing conversation as part of the resilience plan from the outset, not something they figure out after procurement has already committed to a position.

Design for flexibility, not just supply

The longest-term fix is upstream of all of this: getting sourcing and engineering to qualify alternative component equivalents early in the product lifecycle, before a shortage forces the decision under time pressure. This sounds obvious yet gets skipped constantly, usually because qualifying alternates takes engineering time that feels hard to prioritize when the current component supply looks fine. The teams that avoid the worst of an allocation crisis are usually the ones who did this qualification work 18 months before they needed it, not during the crisis itself.

None of this eliminates sub-tier volatility. Memory, storage, and semiconductor allocation cycles are going to keep moving in ways that are hard to predict, and no amount of Tier 1 optimization changes that. What separates the companies that absorb these disruptions from the ones that get knocked off schedule is whether they built visibility, sourcing flexibility, and financing structure at the sub-tier level before they needed it, rather than scrambling to build all three at once when the next allocation cycle turns.

About the author

David Jeng is CEO of Wintec Industries, a company that provides key hardware components, supply chain logistics, and flexible inventory financing solutions to global technology companies.

Today is Thursday, Sept. 24.

1. U.S. financing plan targets Argentine minerals and energy capacity

The U.S. Export-Import Bank plans to finance as much as $7 billion of Argentine critical-minerals and energy projects, potentially supporting equipment purchases from American manufacturers over the next two years.

Reuters  |  Read the original article

2. Mercuria commits $500 million to U.S. strategic mineral inventories

Mercuria pledged $500 million to Project Vault, an initiative supported by the U.S. Export-Import Bank to build critical-mineral inventories that industrial users could access during supply disruptions or market dislocations.

Reuters  |  Read the original article

3. Major shippers pool demand for 2,500 electric heavy trucks

Microsoft, PepsiCo, Ikea and Red Bull joined a coalition ordering 2,500 battery-electric Class 8 trucks, using aggregated demand and leasing to lower deployment costs across major U.S. freight hubs.

Supply Chain Dive  |  Read the original article

4. Tariff period reshapes U.S. import volumes and sourcing patterns

A Descartes Datamyne analysis found U.S. imports fell 4.5% in the year after the April 2025 tariff announcement as sourcing shifted away from China toward Mexico, Vietnam, Taiwan and Southeast Asia.

FreightWaves  |   Read the original article

5. Amazon links Los Angeles imports directly to eastern fulfillment centers

Amazon launched Standard Ocean Express, routing eligible seller inventory through Los Angeles and direct rail to East Coast fulfillment centers; pricing and comparative transit details were not disclosed in the announcement.

Supply Chain Dive  |   Read the original article

6. Old Dominion schedules 4.9% increase on selected LTL tariffs

Old Dominion Freight Line will apply a 4.9% general rate increase to selected standard LTL, cubic-meter and fuel-related tariffs beginning Oct. 5, with actual changes varying by lane and customer.

Supply Chain Dive  |   Read the original article

7. Uniper locks in long-term synthetic jet-fuel supply

Uniper agreed to buy 40,000 metric tons of synthetic aviation fuel annually for more than a decade from Arcadia eFuels’ planned Danish project, with initial deliveries expected in the early 2030s.

Uniper newsroom  |  Read the original article

8. Qualcomm adds MoveIt steward to robotics software portfolio

Qualcomm agreed to acquire PickNik Robotics, steward of the open-source MoveIt framework, seeking closer integration with its Dragonwing platforms while promising continued community governance and support for third-party hardware.

Robotics 24/7  |   Read the original article