Executive takeaways

With AI now entering routine supply chain work, chief supply chain officers (CSCOs) are racing to make their organizations “AI ready.” Yet readiness alone is too limited an ambition.

If CSCOs focus only on adoption, they risk missing the bigger shift already underway: roles are changing, team structures are changing, and the way work gets done is changing with them. Gartner research shows that leaders who take this big picture approach by focusing on AI-human collaboration and its organizational implications outperform laggards who focus primarily on technology adoption.

The pressure to rethink how work gets done in supply chain is only growing, with 88% of supply chain leaders surveyed by Gartner believing it likely or very likely that advancements in agentic AI alone will require new processes for future talent pipelines. That urgency is coming from the top with CEOs looking to leaders across the business to help drive transformation with AI.

That makes AI more than a technology decision. It is a talent, process, and performance issue that sits squarely with supply chain leadership. To capture the full value, CSCOs need to rethink the expectation for their talent, how work is organized and how teams perform through the lens of AI-human collaboration.

To maximize success, CSCOs should prioritize the following three shifts.

1. From “do my job” to “co-evolve with AI”

The first shift is a change in mindset. Many supply chain teams treat AI as a tool that can help people efficiently do existing tasks. However, as AI agents start proposing options, explaining trade-offs, and taking guided action, they can play a larger role in how work gets done. This changes what employees are expected to contribute and how leaders should provide clarity on that expectation.

CSCOs should prepare teams to work with AI as a collaborator, not as a background application. This calls for stronger AI and data literacy and better judgment. Planners need to know when to trust a recommendation from AI and when to push back. Frontline supervisors need to know how to use time saved by automation in ways that raise the quality of decisions, coaching, and execution.

In practice, that could mean a planner spending less time assembling reports and more time resolving exceptions, or a frontline manager using AI-supported insights to coach team members more effectively. The goal is to raise the value of human work as AI takes on more of the routine load.

2. Redefine performance measurement for human–AI teams

As AI becomes part of execution, CSCOs need a broader definition of team performance. Beyond meeting supply chain goals and objectives in the realm of cost and revenue, CSCOs also need to factor in how people and AI work together under real operating conditions.

For example, if an AI agent flags a likely supply disruption earlier than the team would have spotted it on its own, the real measure of success is what happens next. Did the team act quickly, weigh the trade-offs, and make a stronger decision because the signal arrived sooner? That is the kind of intelligence CSCOs need to understand and measure.

3. Embrace agility in workflows and team formations

Many organizations still respond to friction by revisiting job boundaries, process maps or individual KPIs. Those tools have their place, but are not fast enough in supply chains where priorities are changing and decisions are made in real time in response to disruptions. When teams must wait for a formal process review to adjust roles or handoffs, confusion lingers longer than it should, and momentum is lost.

CSCOs need a more adaptive operating rhythm. Core accountability should remain clear, but teams should be able to adjust parts of the workflow as conditions change. Consider a team working from an outdated process design while AI is already predicting demand shifts and triggering earlier replenishment signals. If role clarity is revisited only during a quarterly review, the result is avoidable delay, duplicate effort, and missed value. However, if the team can rework handoffs and regroup around emerging needs, they will be better positioned to keep pace with the business and get more from AI.

Leading through AI-human collaboration

AI will create the most value in supply chains when leaders treat it as an operational issue, not just a technology rollout. That means preparing employees to work alongside AI, redefining what strong team performance looks like, and giving teams more flexibility in how work is organized. CSCOs who make those shifts will be in a stronger position to use AI in ways that improve execution, strengthen decisions, and keep the organization moving as the environment changes.

About the author

Mel Mohamednur is a Director Analyst with Gartner’s Supply Chain Practice. As part of the supply chain talent team, Mel works with CSCOs, head of strategy, and supply chain leaders to navigate talent strategy development. Mel and other Gartner analysts will provide additional insights on human and AI collaboration at the Gartner Supply Chain Symposium/Xpo, taking place May 4-6 in Orlando, FL. Follow news and updates from the conferences on X using #GartnerSC.

Executive takeaways

Despite the enormous effort to meet customers’ expectations, e-retailers often struggle with late deliveries. News reports indicate that the struggle is real. Sure, consumers expect free and fast delivery (Kumar, 2025; Riedl & Mehta, 2026), but final mile operations are expensive and risky. They represent over 50% of e-fulfillment costs and are a key driver of customer satisfaction (Yerapothina, 2025). Furthermore, 87% of shoppers feel that the retailer is fully responsible for delivery delays, regardless of whether the fault lies with the warehouse or the carrier (Convey, 2018).

To address this, retailers typically pour investments into siloed solutions like faster picking bots, better packing materials, or more aggressive carrier contracts. However, these efforts often overlook the entire process flow and, more importantly, the psychological and operational "tug of war" that occurs when managers and workers face mounting order queues.

Our research, recently published in the Journal of Business Logistics (Masorgo et al., 2026), investigates the intricate relationship among order processing (picking, sorting, packing), order delivery, and ultimate on-time performance. By analyzing a dataset of over 10,000 orders from a major e-retailer and conducting interviews with senior operations managers, we identified three critical levers that dictate whether an order arrives on time or falls through the cracks.

1. Early detection is everything

The first step in winning the tug of war is knowing when you are actually losing. Many managers operate in a “visibility vacuum,” where they can see the length of the queue but cannot identify which specific orders within that queue are jeopardizing the delivery promise.

We found a concave relationship between lateness and Order Processing Time (OPT), defined as the proportion of total planned lead time consumed by picking, sorting, and packing. Our data show that orders enter a “danger zone” as soon as internal processing eats up more than 25% of the e-retailer’s total promised lead time.

When processing exceeds this threshold, the pressure shifts downstream to transportation, often leaving carriers with an impossible window. As one operations manager noted: “During the day, pick stations have long lists, then pack stations also have a long line. There is no way to know which orders are about to be late to process them first. We focus on reducing the lines, but the cart with the late order can be stuck way in the back.”

Deeper dive: Elaborating Theory of Swift Even Flow in E-Fulfillment Operations

To counter this, managers must implement monitoring systems that flag orders the moment they hit the critical OPT mark (e.g., 25% in our study), allowing for surgical precision in taking action before the delay becomes irreversible.

2. The art of strategic deprioritization

As internal processing time climbs, a natural tug of war ensues. Managers initially try to compensate for warehouse delays by expediting transportation. They might switch a standard shipment to a premium courier or authorize overtime for drivers.

However, there is an economic and behavioral limit to this recovery effort. Our research found that when OPT exceeds 58% of the e-retailer’s planned lead time, a shift occurs. Managers begin to deprioritize the order.

This is a behavioral response to lost causes. When an order is likely to be late, managers often choose to protect the system’s overall flow rather than wasting expensive resources on an order that looks like it will miss its window. One manager summarized this pragmatism: “Don’t sacrifice the mass to save the few. If an order is late by a day, does it matter if it is two days late?”

Understanding this critical benchmark is vital. Without it, operations often fall into the trap of “re-picking,” or resending items to the picklist because they haven’t reached the transportation bay in time. This creates ghost inventory and overwhelms the staff. By formalizing this threshold (e.g., 58% in our study), companies can stop chasing doomed orders and redirect those resources to ensure that the other orders in the queue stay on track.

3. The complexity lag: Simplicity as a speed lever

The third lever involves the nature of the order itself. It is well-documented that larger, multi-item baskets are harder to pick, but the impact on lateness is compounded by workers’ and managers’ behavior.

Managers and workers naturally gravitate toward less complex orders, such as those containing multiple units of the same SKU or very few items. These orders allow for consistent, rhythmic movements in picking and packing. We found that these simple orders are delivered approximately 8 hours faster than complex ones.

Managers can use this to their advantage by creating dedicated fast-track workflows for low-complexity orders. By clearing the easy wins quickly, they reduce the total volume of the queue, allowing specialized teams to focus on the high-complexity baskets that are more prone to errors and delays.

Breaking the cycle

The friction between warehouse operations and transportation is where the customer experience often falls apart. The battle against lateness cannot be won through brute-force speed; it requires a sophisticated understanding of these operational tipping points.

To move forward, supply chain leaders should:

• Synchronize visibility: Ensure the warehouse management system (WMS) and transportation management system (TMS) share a single lead time clock.
• Empower decisive deprioritization: Document and use critical thresholds to prevent the bullwhip effect of late orders clogging up the system.
• Buffer for complexity: Account for the 8-hour complexity lag when promising delivery windows for multi-item baskets.

By mastering the critical thresholds, e-retailers can stop the internal tug of war and improve on-time performance.

References

Convey. (2018). Last Mile Delivery: What Shoppers Want and How to #SaveRetail. Convey Retrieved from https://www.getconvey.com/press-d-last-mile-delivery-save-retail/

Kumar, N. (2025). Navigating the future of e-commerce logistics: Balancing speed and cost. Supply Chain Management Review https://www.scmr.com/article/navigating-the-future-of-e-commerce-logistics-balancing-speed-and-cost

Masorgo, N., Hoang, T. T. (Jenny), Dobrzykowski, D. D., Bell, J., & Swink, M. (2026). Elaborating Theory of Swift Even Flow in E‐Fulfillment Operations. Journal of Business Logistics, 47(2). 10.1111/jbl.70063

Riedl, P., & Mehta, P. (2026). How autonomous fulfillment is rewriting the rules of supply chain execution. Supply Chain Management Review https://www.scmr.com/article/how-autonomous-fulfillment-is-rewriting-the-rules-of-supply-chain-execution

Yerapothina, S. T. (2025). Unlocking the last mile: A strategic framework for in-store fulfillment. Supply Chain Management Review https://www.scmr.com/article/unlocking-the-last-mile-a-strategic-framework-for-in-store-fulfillment

About the authors

Nicolò Masorgo (Ph.D., University of Arkansas) is an Assistant Professor in the Farmer School of Business at Miami University. Drawing on his order fulfillment and logistics operations industry experience, his main research interest focuses on last-mile delivery operations, service operations, and service supply chain management. His research has been published in several leading journals, including Journal of Business Logistics, International Journal of Physical Distribution and Logistics Management, and Transportation Journal.

Thu Trang Hoang is an empirical supply chain researcher. She graduated from the University of Tennessee, Knoxville, with a PhD in Supply Chain Management. Her research focuses on three main topics: traceability (i.e., food and human trafficking), operational adaptability (i.e., firms’ roles in community relief during disasters), and crowdsourced logistics/e-commerce (i.e., drivers; behaviors under new service/insurance launches, and mathematical modelling).

David D. Dobrzykowski is a Professor of Supply Chain Management and Senior Director of the SCM PhD program at the Walton College of Business, University of Arkansas. His research examines operations and supply chains that feature unique challenges to information processing and the coordination of work processes, such as in healthcare, humanitarian, and sharing economy contexts. He has published in Journal of Business Logistics, Production and Operations Management, Journal of Operations Management, Decision Sciences, Journal of Supply Chain Management, among other leading outlets.

John E. Bell is the Dove Professor of Supply Chain Management at the University of Tennessee.  He holds a doctorate in Management from Auburn University. His research focuses on raw materials, transportation, and sustainable supply chains. He has published over 40 articles in journals such as Journal of Business Logistics, Transportation Journal, and Journal of Operations Management. Prior to joining UT in 2010, Dr. Bell was a career military officer.

Morgan Swink is the Eunice and James L. West Chaired Professor of Supply Chain Management, and Executive Director of the Center for Supply Chain Innovation in the Neeley School of Business, TCU. He teaches and leads research in areas of supply chain management, innovation management, project management, and operations strategy. He has co-authored two supply chain operations text-books, one managerial book on supply chain excellence, and published more than 100 articles in a variety of academic and managerial journals.

Executive takeaways

Despite the advancements in POS (point-of-sales) systems, retail—or notably certain major retailers—still have an inventory accuracy problem. The issue is not due to theft or shrinkage, but according to an insightful December 16, 2024, article on Retail Dive by Daphne Howland, it is due to their use of a draconian method of accounting.

The article states that Nordstrom and Macy’s are moving away from the problematic “retail inventory method” after decades of use, making the shift to “cost accounting.” But based on annual reports (as per the article), retailers Dillard’s, Target, Walmart, Kohl’s, J.C. Penney and Dollar Tree are still using the retail inventory method. One quarter of the National Retail Federation’s top 100 retailers fully rely on the retail inventory method. Other retailers use a partial or hybrid methodology, sometimes due to mergers and acquisitions. 

Created in the 1920s by Malcom McNair, the retail inventory method calculates inventory based on the retail price without counting the inventory. While this was a real time-saver “back in the day” when retail prices probably didn’t change often and before technologies like barcode scanning, the methodology has realistically lost its effectiveness in modern times where retail prices are in a state of flux.

Using the retail inventory method can be like a dog or a cat chasing its tail: because inventory is determined upon price, discounts distort the inventory balance. And how do retailers get rid of excess inventory? Markdowns and discounts create more distortions in the calculation. And because buyers buy based on inventory count, a distorted inventory count has a knock-on effect on the buy or re-buy quantity, worsening the problem. 

Is the shrink issue that’s been in the news really as tied to theft as it has been made out? It is possible that a notable portion of this shrink is actually due to internal issues including administrative mistakes and process errors. 

If inventory counts are not accurate, wouldn’t this also affect forecasts if the retailer offered these to their vendors? What about the accuracy of product activity (sales) data … could this information be compromised as well? Retailers continually blame vendors for their inventory woes, but if retailers cannot provide accurate forecast and inventory data to their vendors (as well as their own personnel), aren’t the retailers as much or even more to blame for their inventory problems, whether it’s too much or too little or too late? 

It's in everyone’s best interest to ensure that inventory is always available when and where the consumer wants it. That’s part of achieving the perfect order. But if the data isn’t accurate, it’s going to have an affect on execution.        

Executive takeaways

A planner walks into a morning operations meeting with three dashboards open. One shows excess inventory, another signals potential stockouts, and a third highlights rising logistics costs. Each is accurate. None leads to a clear decision. The first half of the meeting is spent reconciling numbers. The decision is delayed or escalated. Everyone has the data, but no one owns the decision.

This situation is no longer unusual. Over the past few years, companies have invested heavily in control towers, real-time visibility tools, predictive analytics, and AI. The expectation was straightforward: more data would enable faster and better decisions. Instead, many organizations are experiencing slower decision cycles, more alignment meetings, and increased hesitation. The problem is no longer data scarcity. It is how decisions are made in a data-rich environment. Supply chains are becoming data-rich but not decision-ready.

Why more data is not leading to better decisions

This issue has become more visible as supply chains operate under continuous disruption, where volatility is no longer episodic but structural. In this environment, speed matters as much as accuracy, because a delayed response to a disruption can quickly erase the value of even the most accurate forecast.

At the same time, expectations around “data-driven decisions” have shifted. Leaders are now expected to justify actions with analytics, not just experience. While this improves transparency, it also introduces friction, as decisions are validated, rechecked, and aligned across functions before anyone commits to a course of action. Many organizations now face what a recent industry report describes as an “insight-to-action gap,” where the ability to generate insights exceeds the ability to act on them in time. The underlying issue is not data quality but that decision processes have not evolved at the same pace as data capabilities.

Where decisions break down

Across organizations, similar patterns appear. First, metrics often conflict. Procurement may focus on cost, operations on throughput, and customer‑facing teams on service levels. Each function works with accurate data, but there is no clear rule for which metric takes priority when trade-offs arise. Targets are met locally while the end-to-end decision stalls, and meetings are spent defending metrics rather than choosing a path. Teams optimize locally, but the organization cannot decide globally.

Second, decision ownership is often unclear. Data is widely accessible, but authority is not. Multiple teams can see the same issue, yet no one is clearly accountable for acting on it. Planners hesitate, managers forward issues to peers, and problems travel “up and sideways” until someone senior feels compelled to step in. This creates hesitation, frequent escalation, and a culture where analysis moves faster than decisions. Issues move faster than authority.

Third, more data introduces more validation. Leaders seek confirmation across systems before committing to a decision, asking for one more report, another scenario, or a cross‑check from a different team. While this reduces risk in isolated cases, it slows the organization. Speed is traded for certainty, even when certainty is unattainable. Over time, people learn that it is safer to ask for more data than to decide with the information already available. These patterns reinforce each other. More visibility exposes differences in priorities. Differences require alignment. Without clear ownership, alignment becomes escalated. The result is slower decision‑making despite better information.

The issue can be summarized simply:

When decisions catch up to data

Some organizations are addressing this gap by designing decisions first, and data second. UPS’s Harmonized Enterprise Analytics Tool (HEAT) is often cited for its analytics capabilities. Its impact, however, comes from how it supports specific operational decisions such as routing and capacity allocation. HEAT ingests more than a billion data points per day to create a single view of network performance and feeds that into planning and management routines that adjust how packages move through the network in near real time. Rather than presenting all available data, the platform emphasizes what matters most for those decisions and embeds it in day‑to‑day operations.

PepsiCo took a similar approach with its sales intelligence platform. Instead of building a broad analytics hub, the company focused on a single decision: predicting and preventing out‑of‑stocks at the store level. Its AI‑driven demand forecasting, developed with partners such as TAZI, has achieved about 98% accuracy for most products and reduced truck stock‑outs by roughly 4%, while improving order size and product mix on delivery routes. By tying analytics to a concrete decision and playbook, PepsiCo made it easier for teams to act quickly rather than simply observe more data.

Pfizer’s Global Supply Digital Operations Center also illustrates this shift. The DOC functions as a virtual cockpit for manufacturing and supply, providing a shared, end‑to‑end view of operational performance across sites. Pfizer reports that the DOC has helped reduce cycle time in some areas and, more importantly, has “transformed how manufacturing colleagues collaborate and make decisions,” enabling teams to predict issues before they occur and adjust in real time. The emphasis is not on adding more dashboards, but on speeding and coordinating interventions where they matter most. In each case, technology is important, but the design principle is more important. Data is organized around decisions, not the other way around.

Making supply chains decision-ready

For supply chain leaders, the key question is no longer how to improve visibility. It is how to improve decision speed and clarity. Figure 1 shows a supply chain decision-making pathway. A practical starting point is to identify a small set of critical decisions where speed matters most, such as disruption response, allocation across channels, or supplier adjustments. For each decision, three elements should be clearly defined: who owns the decision, how quickly it needs to be made, and which metrics take priority when trade-offs occur. Writing these down and socializing them turns vague “data-driven” expectations into an explicit playbook.

Leaders should also examine the dashboards and tools used in operations. Each should be tied to a specific decision and cadence. If a dashboard does not clearly answer “what action should be taken,” it is adding noise rather than value. In many organizations, simply retiring or redesigning a few widely used dashboards removes friction and reduces time spent reconciling numbers.

Finally, analytics efforts should be aligned with decision cycles. Instead of building general-purpose tools, organizations should design analytics to support specific decisions at defined intervals. This shift forces clarity about what information is necessary and what is sufficient to act, and it helps analytics teams measure success in terms of faster, better decisions rather than the number of reports delivered.

Figure 1: Supply Chain Decision Making Pathway

From data-rich to decision-ready

Most supply chains have already solved the problem of visibility. The next challenge is execution. Organizations that continue to invest primarily in data capabilities may see diminishing returns. Those that focus on how decisions are made will move faster and respond more effectively to disruption. The advantage is no longer having more data. It is the ability to act on it. Most supply chains do not lack data. They lack clarity on who decides, how fast, and based on which signals.

About the author

Corrine Chen is an educator, researcher, and former industry executive with over a decade of hands-on experience in supply chain management, procurement, and innovation. She teaches supply chain management courses at the University of Nebraska Omaha. Corrine’s work bridges academia and practice, with published research, applied projects, and a passion for empowering the next generation of supply chain professionals. She can be reached at peonyhill@yahoo.ca.

Executive takeaways

It’s Tuesday morning. A VP of supply chain opens her inbox and sees a carrier notice about port congestion in Los Angeles, a broker email about updated CBP documentation requirements, three headlines about tariff developments, and a supplier email flagging delays out of Vietnam. All of it arrived overnight, but none of it tells her which, if any, require action before her 9 a.m. call with the CFO.

Brad McDougle

That’s not an unusual morning. It’s a typical one.

Importers don’t lack information. They get carrier updates, broker notices, tariff headlines, customs guidance, supplier emails, and market reporting all day long. What many still lack is a clear way to decide which outside developments actually matter to their sourcing profile, trade lanes, customer commitment and margin.

Importers aren’t operating in a stable environment with the occasional disruption anymore. This is not getting easier to manage. Thomson Reuters reported in its 2026 Global Trade Report that 72% of trade professionals identified U.S. tariff volatility as the most impactful regulatory change they face, up from 41% a year earlier. Supply chain management rose to 68% as a top strategic priority, up from 35% the year before. Those are real changes in what supply chain teams now have to watch every day.

The usual response is to throw more data at the problem. But more information doesn’t automatically lead to better judgment. Dashboards can pull signals into one place and AI tools can process them faster, but neither can tell a company what a specific development actually means for its business. In many organizations, someone still has to make that call. Large multinationals may have dedicated risk, trade, or intelligence teams focused on that work full time. Many mid-market importers don’t.

Most of the time, external risk gets pushed onto trade compliance, logistics, procurement, or finance as an add-on responsibility, usually on top of a full-time role that’s already demanding. The result is predictable. Signals get picked up in pieces, without focused attention or anyone clearly responsible for deciding what needs escalation. By the time a development is recognized as urgent, they may have already missed the window to respond.

Trade policy makes the problem easy to see. The recent tariff turmoil shows why. Importers aren’t dealing with a single tariff announcement followed by a clean downstream response. They’re dealing with a constant stream of policy changes, legal challenges, agency guidance, retaliatory measures, exemption questions, and sourcing implications. What matters is whether a development changes landed cost assumptions, creates exposure on open purchase orders, affects a key supplier country, or needs immediate attention.

Mistakes here have become expensive. An ECB study cited by Reuters in March 2026 found that U.S. consumers and importers are absorbing most of the financial hit from tariffs, not foreign exporters. Thomson Reuters reported that 76% of trade professionals now believe the current tariff environment reflects a permanent policy shift rather than a short-term problem. According to the same report, 65% of trade professionals are already changing sourcing patterns, while most are renegotiating supplier contracts or moving manufacturing closer to home. Those are major operating decisions made under uncertainty, which makes good judgment even more important.

Customs and enforcement create a different kind of exposure because the signals rarely look urgent until they affect a real shipment. A common example looks like this: CBP updates its documentation expectations for a particular product category. The notice goes to the compliance team, gets logged, and gets marked as something to monitor. Logistics isn’t looped in. A purchase order is already in transit. The shipment arrives and gets flagged for additional documentation the broker wasn’t prepared for. The release is delayed a week. Now the product that was supposed to ship on Friday is sitting there, operations is scrambling, the customer is asking questions, and finance is trying to calculate the cost of the delay. The compliance team saw the notice. What was missing was someone clearly responsible for deciding what needed to happen next.

Supply Chain Management Review noted in January 2026 that trade compliance has moved from a back-office function to a strategically and legally exposed leadership responsibility. That shift shows up in enforcement activity, but it also shows up in who gets asked to explain what happened and why it wasn’t caught earlier.

In logistics, the problem looks a little different. Port advisories, carrier notices, labor developments, severe weather, and chokepoint disruptions often look routine on their own. A notice about congestion at the Port of Miami reads the same way whether a company has three containers sitting at the terminal waiting for chassis or none. A blank sailing announcement from a major carrier looks like background noise until someone checks it against the replenishment schedule for a product line running at six days of inventory cover. The alert was accurate. It just didn’t mean much without the right operating context.

The difference isn’t always how much information a company has. It’s whether anyone is clearly responsible for watching the environment, filtering for relevance, applying judgment, and getting the right signal to the right person early enough for it to matter.

For many importers, that responsibility is still scattered. The carrier sent the notice. The agency published the update. Someone saw the headline. What was missing was not access to the data. It was a lack of clarity about who needed to review it, who needed to raise it, and who needed to respond.

About the author

Brad McDougle is the founder of Import Risk Intelligence. His work focuses on trade policy, customs and enforcement developments, and logistics disruptions affecting U.S. importers. He previously served as a DSS Special Agent and has also operated a U.S. importing business.

Executive takeaways

Airlines manage billions in spend across safety-critical aircraft components, FAA-regulated maintenance services, airport operations, and enterprise technology. Yet many organizations still operate with fragmented processes and loosely connected systems. This limits visibility into how spend is planned, committed, and governed.

Anshul Bansal

Even with modern source-to-pay platforms, insufficient integration with contracts, projects, and financial systems can slow improvements in compliance, spend transparency, and working capital management. Closing this gap requires viewing procurement as an integrated capability within the broader enterprise operating model rather than a standalone system upgrade.

The airline procurement reality

Airlines manage a uniquely complex mix of indirect and direct spend. While indirect categories such as IT, HR, facilities, real estate, marketing, and legal are often governed through enterprise procurement systems, large volumes of operational spend, such as aircraft parts procurement, ground services, catering, and airport operations, have historically operated outside structured S2P processes.

Historically, much of this operational spend bypassed structured procurement. Non-PO transactions were common, approvals occurred outside formal systems, and invoices arrived after services were rendered. Spend data was fragmented across ERPs, legacy maintenance systems, spreadsheets, and emails. Even when strong commercial contracts were negotiated, they were frequently disconnected from execution, stored in offline repositories, unmanaged SharePoint sites, or paper files, with no systematic enforcement at the point of spend.

Compounding the challenge, many aircraft maintenance and parts systems are FAA-regulated, decades-old platforms that are mission-critical and difficult to replace. Rather than attempting wholesale replacement, leading transformations have focused on synchronizing these environments with modern S2P platforms, preserving operational integrity while enabling enterprise governance and visibility.

Integrating direct spend without disruption

Integrating operational spend into S2P proved to be as much an organizational challenge as a technical one. Maintenance and airport operations teams prioritize speed and safety, and any perceived friction can undermine adoption.

Successful programs respected these realities by allowing teams to continue working in certified systems while synchronizing contracts, commitments, and financial controls end-to-end. This approach embedded governance without disrupting execution, creating a compliant and auditable procurement environment for direct spend at enterprise scale.

Operationalizing contracts and suppliers

A major source of value unlock came from operationalizing contracts directly within procurement workflows. Thousands of contracts and active projects were cleansed, digitized, and embedded into transactional workflows.

This shift converts unmanaged spend into contract-enabled purchasing, ensuring pricing, terms, and regulatory requirements are applied automatically at requisition and invoice stages. Standardized supplier onboarding and catalog-driven buying further reduce spend leakage and accelerate supplier enablement. As a result, spend under management has increased from around 40% to 75% to 80% in large-scale programs.

Enable enterprise IT spend

Enterprise IT procurement is complex, with a diverse user base that includes corporate staff, airport personnel, pilots, and flight crews. Some programs have introduced reusable integration connectors between major SaaS platforms to orchestrate IT hardware procurement directly into S2P workflows. These connectors standardize processes for laptops, tablets, mobile devices, and peripherals. Unlike one-off integrations, they can be reused across the enterprise and industry. This improves usability, keeps auditability intact, and ensures operational and financial controls are aligned without disrupting daily workflows.

Measurable outcomes

Across large airline S2P programs, results included:

• Spend under management increased from roughly 40% to 75% to 80%
• Supplier onboarding cycle times dropped from 20–25 days to around 7 days
• Invoice exception rates decreased significantly due to upstream enforcement
• Procurement and payment cycles became faster, improving operational responsiveness
• Multiple legacy procurement and finance systems were sunset, reducing overhead and improving data accuracy

Operational categories such as catering, cleaning, baggage handling, and ground services also saw productivity gains. Suppliers gained better visibility into contract terms, invoice status, and payment schedules, reducing exceptions and improving collaboration.

Organizational realities and trade-offs

These transformations are not executed without challenges. Integrating regulated legacy platforms takes time and careful planning. Resistance can occur when teams are used to legacy systems. Success often requires phased rollouts, role-based training, and close collaboration between procurement, finance, IT, and operational teams. When managed well, the long-term benefits in visibility, compliance, and enterprise agility outweigh the upfront effort.

Procurement as an operating system

Source-to-pay transformation is no longer just a back-office modernization initiative. It has become a strategic enabler of operational resilience, compliance, and value realization. Integrating legacy and modern systems, direct and indirect spend, and supplier and internal workflows shows that procurement can directly influence financial performance, working capital efficiency, and operational agility.

Measurable results, from higher spend under management to faster supplier onboarding and improved compliance demonstrate that procurement can move from a transactional function to a core driver of enterprise-wide performance. For airlines facing margin pressures, complex regulations, and distributed operations, mature S2P models are essential infrastructure for sustaining growth, reducing risk, and capturing the full value of negotiated contracts.

Author disclosure

The author holds over 18+ years of technology consulting experience and has held senior technical and delivery leadership roles on multiple large-scale airline Source-to-Pay transformation programs and continues to advise airline clients on procurement and technology modernization initiatives. This article reflects industry insights and observed outcomes rather than personal opinion.

About the author

Anshul Bansal is a technology consulting leader at Accenture LLP with 18+ years of experience in Big 4 firms for designing and delivering large-scale Procurement, Finance, and Supply chain transformation programs. He specializes in implementing Source-to-Pay (S2P) and spend management transformations, helping organizations translate strategy into executed outcomes.

Throughout his career at leading consulting firms, including Accenture and Deloitte Consulting, he has served as a trusted advisor to C-suite executives, guiding complex enterprise transformations spanning sourcing, supplier risk management, procurement operations, and financial systems integration. He works closely with executive sponsors and functional leaders to modernize operating models, integrate digital platforms, and embed new capabilities into day-to-day execution. From experience working with Fortune 500 clients across several industries, he has led one of the industry first transformation for US major airline carrier and had further replicated to other major airline carrier. He serves as an industry expert advising several airline clients on leading practices for end to end process re-engineering and also define solutions to connect with decades old complex direct spend procurement applications.  

His recent work focuses on technology reinvention through analytics, automation, and AI, with a pragmatic emphasis on breaking down data silos, reducing manual workarounds, and enabling more resilient and adaptive supply chain operations.

Executive takeaways

Warehouse automation is no longer a differentiator—it is a prerequisite for operating at scale. Rising labor costs, volatile demand patterns, tighter delivery expectations, and increasing SKU complexity have forced organizations to accelerate investments in goods-to-person systems, automated storage, high-speed sortation, and automated packing. Yet as automation density increases, many enterprises discover that technology selection alone does not translate into predictable performance or sustainable return on investment.

The primary challenge has shifted from what automation to deploy to how automation is orchestrated. Modern warehouses now operate as complex systems of systems, where SaaS-based warehouse management systems, automation execution platforms, and on-premises control software must function as a cohesive ecosystem. In these environments, the success or failure of automation initiatives is determined less by mechanical capability and more by the underlying software architecture that coordinates decisions, execution, and exception handling in real time.

A common misconception in automation programs is the assumption that vendor-provided systems will naturally integrate “out of the box.” In practice, each platform introduces its own data models, service-level assumptions, and failure behaviors. SaaS WMS platforms, while offering faster deployment and continuous innovation, are intentionally constrained to configuration over customization, limiting direct control over execution logic. As a result, the responsibility for ensuring scalability, resilience, and predictable behavior increasingly falls on the customer’s integration and orchestration design.

This practitioner-led case study examines the software architecture behind a highly automated, greenfield warehouse implemented using commercially available SaaS and automation platforms. Rather than describing a single company’s solution, the article presents a generalized, reusable reference architecture that reframes the SaaS WMS as an orchestration layer rather than a control system. The intent is to provide supply chain and technology leaders with a practical blueprint for reducing automation risk, improving time-to-value, and future-proofing warehouse platforms in an era of rapid change.

Why businesses need automation-first warehouses

Warehouse automation has shifted from a long-term aspiration to a near-term business necessity. Rising labor costs, fluctuating demand, tighter delivery windows, and increasing SKU complexity have fundamentally changed the economics of warehouse operations. Automation-first warehouse designs are no longer driven solely by efficiency gains, but by the need to create resilient, scalable fulfillment platforms that can operate predictably under peak conditions.

At the inbound edge, automation enables no-touch or low-touch receiving and storage, reducing manual handling while improving accuracy and throughput. Automated putaway, buffering, and goods-to-person systems allow inventory to be stored densely and retrieved efficiently, enabling better utilization of vertical and horizontal space. This is particularly critical as facilities grow larger and land and construction costs continue to rise. Automation-first designs allow organizations to maximize usable storage capacity without proportionally increasing labor or footprint.

On the outbound side, automation enables faster and more consistent fulfillment by reducing travel time, sequencing work intelligently, and decoupling picking, packing, and shipping activities. Automated sortation, buffering, and packing systems support high-throughput order processing while maintaining accuracy. These capabilities are increasingly essential to meet same-day or next-day delivery expectations without relying on unsustainable labor models.

Automation also plays a direct role in shipping cost optimization. By consolidating multiple units across orders and intelligently combining items stored both within and outside automated zones, organizations can reduce carton count, optimize packaging, and lower transportation costs. The business value of automation is therefore not limited to labor savings, but extends across storage efficiency, fulfillment speed, and transportation economics.

However, while the business case for automation-first warehouses is compelling, many organizations underestimate the complexity of making these environments function as a cohesive system. That complexity does not reside primarily in the automation equipment itself, but in the software architecture required to orchestrate it.

Why automation-first warehouses need a software architecture model

Modern warehouses are increasingly defined not by individual automation technologies, but by the software platforms that orchestrate them. As automation density increases, modern warehouses now operate as systems of systems, with multiple automation technologies coexisting within a single facility. Goods-to-person grids, sorters, buffers, conveyors, and automated packing systems must work alongside traditional pallet, case, and manual picking areas. These mixed environments introduce real-time dependencies across workflows that were previously isolated, requiring precise coordination between automated and human-driven processes.

This complexity is further amplified by the reality of SaaS-based warehouse management systems. While SaaS WMS platforms offer faster deployment and continuous innovation, they prioritize configuration over customization and provide limited visibility into internal processing. As a result, organizations have less control over execution behavior and must rely heavily on the quality of their integration architecture to meet performance and reliability expectations. In automation-first environments, integration becomes the primary mechanism through which operational intent is expressed and enforced.

A common misconception in automation programs is the assumption of “out-of-the-box integration.” Vendors often imply that their systems will naturally interoperate, yet in practice, each platform introduces its own data models, service-level assumptions, and failure modes. Message timing, state ownership, error handling, and retry behavior vary significantly across systems. The responsibility for reconciling these differences — and for ensuring that automation behaves predictably at scale — ultimately falls on the customer.

This gap between expectation and reality leads many implementations to suffer from bottlenecks, timeouts, and operational workarounds that erode the value of automation. When software architecture is treated as an afterthought, automation becomes brittle and difficult to evolve. Conversely, when a clear software architecture model is established upfront, organizations gain resilience, scalability, and the ability to extend automation capabilities without destabilizing operations.

For these reasons, automation-first warehouses require more than advanced equipment or feature-rich platforms. They require a deliberate, software-first architecture model that defines how systems interact, how decisions are orchestrated, and how failures are managed. The following sections outline the architectural principles that underpin such a model and form the foundation for a scalable, automation-first warehouse software platform.

Design principles for an automation-first warehouse software platform

Automation-first warehouses demand a fundamentally different software mindset than traditional, labor-centric facilities. The following design principles form the foundation of a reusable architecture model for orchestrating modern, highly automated warehouse environments. These principles are not tied to specific products or vendors; rather, they define the architectural behaviors required for scale, resilience, and long-term adaptability.

Orchestration over direct control

In an automation-first warehouse, the WMS must act as a system of orchestration, not as a controller of physical equipment. Business intent—such as prioritizing orders, releasing work, or handling exceptions—should be expressed at the orchestration layer, while execution systems manage the mechanics of how that intent is fulfilled. This separation prevents business logic from becoming tightly coupled to specific automation technologies and allows execution systems to evolve independently without destabilizing core warehouse operations.

Event-driven by default

Automation environments generate high volumes of state changes that must be processed in near real time. An event-driven integration model is therefore essential. Systems should communicate through well-defined business events rather than synchronous, chatty interactions. This approach improves throughput, reduces latency sensitivity, and enables systems to absorb spikes in volume without cascading failures. Event-driven design also provides a natural foundation for replay, recovery, and auditability.

Loose coupling across vendor platforms

Automation-first warehouses are inherently multi-vendor ecosystems. Each platform introduces its own lifecycle, release cadence, and operational assumptions. The architecture must deliberately minimize coupling between systems by enforcing clear integration contracts, isolating state ownership, and avoiding shared internal dependencies. Loose coupling enables components to be replaced, upgraded, or expanded without requiring wholesale redesign of the warehouse platform.

SLA-aware interactions

Not all warehouse interactions are equal. Some flows, such as induction or packing confirmation, are highly latency-sensitive, while others can tolerate asynchronous processing. The architecture must explicitly distinguish between these interaction types and apply appropriate integration patterns accordingly. Designing with service-level awareness ensures that automation systems meet operational expectations without over-engineering low-risk flows or introducing unnecessary synchronization.

Failure as a normal operating condition

In highly automated environments, failures are inevitable—network interruptions, equipment faults, and partial outages are part of daily operations. The software architecture must assume failure as a normal condition and provide built-in mechanisms for retry, fallback, and manual intervention. Clear ownership of error states and recovery paths prevents small issues from escalating into systemic disruptions and allows operations to continue safely during degraded conditions.

Policy-driven, not hard-coded behavior

Automation-first warehouses require flexibility to adapt to changing volumes, fulfillment strategies, and business priorities. Behavioral decisions—such as routing, prioritization, or buffering strategies—should be governed by configurable policies rather than embedded logic. Policy-driven design enables organizations to adjust operations without invasive system changes and supports continuous improvement over time.

Scalability through architecture, not customization

Scalability in automation-first warehouses is achieved through architectural discipline rather than extensive customization. By relying on standardized integration patterns, clear system boundaries, and event-based communication, the platform can scale horizontally across throughput, facilities, and automation types.

This approach reduces technical debt and preserves the benefits of SaaS-based systems while supporting complex automation scenarios. These design principles establish the foundation for a modern, automation-first warehouse software platform. The next section applies these principles to a reference software architecture model, illustrating how SaaS WMS platforms, execution systems, and automation technologies can be orchestrated as a cohesive, scalable ecosystem.

The strategic blueprint: The automation-first warehouse

Architecture: From orchestration to execution

Modernizing the warehouse requires shifting focus from individual mechanical components to a unified software and integration orchestration layer. This reference architecture serves as a template for a scalable, resilient ecosystem that minimizes the “technical debt” often associated with automation.

Reference Architecture for an Automation-First Warehouse Software Platform

This reference architecture illustrates how a modern, automation-first warehouse can be structured as a layered, software-driven ecosystem rather than a collection of tightly coupled systems. Each layer plays a distinct role in translating business intent into physical execution while preserving scalability, resilience, and vendor flexibility.

1. Orchestration layer: SaaS warehouse management system (WMS)

At the top of the architecture sits a SaaS-based WMS, serving as the system of orchestration for the warehouse network. Rather than directly controlling automation equipment, the WMS defines what work should be performed—order priorities, inventory allocation, and fulfillment policies—while delegating how that work is executed to downstream systems.

This separation enables executives to scale capacity, introduce new automation technologies, or modify fulfillment strategies without destabilizing core warehouse operations. By operating as a cloud-native orchestration hub, the WMS provides enterprise-wide visibility while supporting rapid deployment across facilities.

2. Integration layer: Event-driven, vendor-neutral connectivity

Beneath the WMS, an event-driven integration layer connects planning systems, execution platforms, and automation software through well-defined business events. This layer replaces rigid point-to-point integrations with asynchronous communication, allowing systems to operate independently while remaining coordinated.

By enforcing vendor-neutral contracts and isolating state ownership, the integration layer absorbs variability in vendor behavior, reduces latency sensitivity, and enables the platform to scale throughput without cascading failures. This design is foundational to operating automation-first warehouses under peak conditions.

3. Execution layer: Warehouse execution and control systems

Closer to the physical environment, warehouse execution systems (WES) and warehouse control systems (WCS) translate orchestration intent into machine-level actions. These systems manage sequencing, buffering, and real-time coordination across automation assets such as goods-to-person systems, sorters, conveyors, and automated packing.

Locating execution logic near the equipment improves operational resilience, allowing facilities to continue operating during transient cloud or network disruptions while maintaining alignment with upstream orchestration decisions.

4. Intelligence layer: Analytics, optimization, and AI

Surrounding the core operational layers is an intelligence layer composed of analytics platforms and optional AI/ML optimizers. This layer transforms operational data into predictive insights, enabling organizations to optimize labor, throughput, and inventory flow proactively rather than reactively.

By decoupling optimization from execution, the architecture allows advanced decision-making capabilities to evolve independently, future-proofing the warehouse as fulfillment complexity continues to increase.

Takeaways

• Automation success is an architecture problem, not a hardware problem.

As automation density increases, the primary risk shifts from equipment performance to software orchestration and integration design.

• The SaaS WMS must act as an orchestrator, not a controller.

Treating the WMS as a coordination layer—rather than direct automation control—improves scalability, resilience, and vendor flexibility.

• Event-driven integration is foundational, not optional.

High-throughput, automation-first warehouses require asynchronous, event-based communication to absorb volume spikes and avoid cascading failures.

• Vendor neutrality preserves long-term strategic flexibility.

Loosely coupled architectures reduce dependency on individual automation providers and simplify future expansion or replacement.

• Future-proofing requires policy-driven design.

Encoding operational behavior through configurable policies—rather than hard-coded logic—enables continuous adaptation as volumes, channels, and fulfillment strategies evolve.

• Reframing the WMS as a cloud-native orchestration hub shifts automation investment from fixed CAPEX to scalable OPEX, improving responsiveness to seasonal demand.

Implications for supply chain & technology leaders

For executives overseeing large-scale warehouse automation investments, the implications are clear. Automation-first fulfillment strategies must be governed as software platform programs, not as collections of isolated equipment deployments. This requires elevating software architecture, integration design, and orchestration ownership to the same level of executive attention traditionally given to mechanical automation and facility design.

Leaders should ensure that SaaS WMS platforms are positioned as systems of orchestration, with clear execution boundaries defined between planning, integration, and automation control layers. Investment decisions should prioritize event-driven, vendor-neutral integration capabilities that preserve long-term flexibility and reduce dependency on individual automation providers. Finally, organizations must adopt policy-driven operating models that allow fulfillment behavior to evolve without repeated system rework, enabling automation platforms to scale sustainably as volume, channels, and customer expectations continue to change.

About the author

Muruganandham Kalimuthu is a principal engineer and software architect at a major U.S. retail company, with experience designing both custom, in-house warehouse management systems and large-scale, SaaS-based WMS and automation platforms. He has led greenfield automation initiatives as well as brownfield modernization programs involving legacy warehouse stacks and phased SaaS adoption across high-volume retail fulfillment networks.

Executive takeaways

There is a pattern buried inside every major technological revolution that operations leaders rarely discuss because it takes us away from the exhilarating feeling of having found the silver bullet to organizational bottlenecks. Every time a transformative technology matures and becomes broadly accessible, organizations do not benefit equally. The ones that fall behind are rarely the ones who failed to adopt the technology. They are the ones who adopted it without understanding where the constraint moved.

Niraj Jha

This is the Law of Constraint Migration. And it is playing out right now with AI, at a speed that will separate this decade's operational winners from its casualties.

The pattern, three times over

We can study the transition from horse-drawn transport to the internal combustion engine in the early 20th century. Before the IC engine, the binding constraint on freight and logistics was biological. Success depended on an intricate operational knowledge: the care and conditioning of horses, the management of stables and feed supply chains, the scheduling of rest cycles, the expertise to read an animal’s health before a long haul. Fleet operators who mastered that knowledge had a genuine competitive advantage. Then the IC engine arrived and made all of it irrelevant almost overnight. The constraint did not disappear. It migrated. The new operational imperatives were fueling infrastructure, mechanical reliability, clutch cables and drive trains and the staffing of skilled mechanics, route planning around fuel stops rather than water troughs. The transport operators who thrived were not simply the ones who bought trucks earliest. They were the ones who recognized that the knowledge defining success had fundamentally changed, and rebuilt their organizations around the new constraint before their competitors understood what had shifted.

The same pattern repeated with electrification. By the early 20th century, electricity was widely available and cheap. Yet factory productivity did not surge immediately. Historians of technology, most notably Paul David in his work on the dynamo and the computer, documented the lag. Factories that simply replaced their existing power arrangements with electric motors gained little. The constraint had migrated from energy generation to factory layout, workflow design, and the organizational logic of production itself. The productivity revolution came only when manufacturers rebuilt their operations from the floor up around what electricity actually made possible.

Then computing. By the 1990s, computing power was commoditized. Hardware was cheap. Software was available. Yet research consistently showed productivity paradoxes across industries. The organizations extracting full value were not the heaviest technology spenders. They were the ones that changed how decisions were made, how information flowed, and how humans worked alongside machines. The constraint had migrated from processing power to organizational capability and process design.

Three revolutions. Three migrations. Each time, the technology became a commodity faster than most organizations could adapt to where the new constraint had landed.

Where AI is moving the constraint now

AI in supply chain and operations is following this exact pattern, but the migration is happening faster than any prior cycle, and the destination is less obvious.

The surface-level constraint AI is dissolving is clear: the time and cost of analysis. Tasks that required a team of analysts working for days, demand sensing, supplier risk scoring, routing optimization, anomaly detection across thousands of SKUs, can now run continuously and autonomously. That constraint is effectively gone for any organization with the infrastructure to deploy modern AI tools.

But the constraint has not disappeared. It has migrated to three places most organizations are not yet looking.

The first is data architecture. AI does not create insight from noise. It amplifies whatever signal exists in the underlying data. Organizations with fragmented systems, inconsistent master data, and siloed operational records will find that AI accelerates their existing dysfunction as readily as it accelerates good decision-making. The constraint has moved from analytical capacity to data quality and governance, and most operations functions still treat data infrastructure as an IT problem rather than a strategic one.

The second is decision rights. Autonomous systems make autonomous decisions. When AI reroutes a shipment, activates a secondary supplier, or adjusts a production schedule without human initiation, the question of who owns that decision, and who is accountable when it is wrong, becomes operationally critical. Organizations that deploy AI without restructuring their decision rights frameworks will discover this constraint the hard way, usually during a high-stakes disruption when accountability is suddenly important.

The third is the capability of the people left in the loop. As AI absorbs routine analytical and transactional work, the humans remaining in operational roles are being asked to do something harder: exercise judgment on novel situations the model has never seen, interpret outputs that arrive without full explanatory context, and make calls that sit at the edge of the system’s confidence. The constraint has migrated from execution capacity to human judgment quality, and most organizations are not investing in that capability with anything like the urgency they are investing in the AI deployment itself.

Why equal access does not produce equal outcomes

This is the part of the AI conversation that does not appear in vendor presentations. When a technology matures and becomes broadly accessible, the naive assumption is that the playing field levels. Everyone has access to the same tools, so outcomes should converge.

They do not converge. They diverge. Because the constraint has migrated, and organizations differ enormously in their readiness at the new constraint location.

Two companies in the same industry can deploy the same AI platform, with the same budget, at the same time, and produce radically different results. The difference will not be the algorithm. It will be the quality of the data feeding it, the clarity of the decision frameworks governing it, and the depth of human judgment available to override it when the situation demands. Those factors are not created in a software deployment. They are built over years, through deliberate organizational investment, and most leaders are not yet treating them as the urgent operational priorities they are.

The organizations that extracted full value from electrification did not simply buy electric motors. They spent a decade redesigning their factories. The organizations that will extract full value from AI will not simply buy platforms. They will spend years redesigning the organizational substrate that AI operates within.

What operations leaders should do differently starting now

The practical implication is not to slow down AI adoption. It is to invest in parallel, with equal urgency, in the three places the constraint has migrated.

On data architecture: treat data quality as a supply chain input, not an IT project. Map the data flows that your AI systems will depend on the same way you would map a supplier network. Identify the single points of failure. Understand where the signal degrades. This is infrastructure investment, and it needs to appear on the capital plan accordingly.

On decision rights: before deploying autonomous systems, define explicitly which decisions they own, which they inform, and which remain with humans regardless of what the model recommends. Build the exception protocols before you need them, not during a crisis. The time pressure of a disruption is the worst possible moment to be clarifying accountability.

On human judgment: identify the roles in your operation where AI will concentrate decision complexity rather than reduce it. These are the people who will be asked to exercise judgment on novel, high-stakes situations with AI-generated context they may not fully understand. Invest in their capability development with the same intentionality you are bringing to the technology deployment itself.

The constraint has always moved

There is a version of the AI conversation that is essentially triumphalist. The technology is powerful, adoption is accelerating, and the organizations moving fastest will win. That version is incomplete.

The more accurate version, the one that the history of industrial transformation actually supports, is that the technology is powerful, the constraint is moving, and the organizations that will win are the ones that figure out where it landed before their competitors do.

The horse did not end the constraint of distance. It relocated it, until the IC engine moved it again. Electrification did not end scarcity. It relocated it. AI will not end operational constraint. It will relocate it, to data quality, to decision architecture, to human judgment at the edge of algorithmic confidence.

The leaders who understand this will not simply be early adopters. They will be the ones who invest in the right things at the right time, because they were looking in the right place when everyone else was still celebrating the deployment.

About the author

Niraj Jha is Senior Director of Logistics at Niagara Bottling, the largest privately held beverage company in the United States, where he oversees a network of manufacturing plants, third-party logistics providers, and the deployment of AI across supply chain operations. He is the author of From Engines to Algorithms and is an avid writer, sharing his ideas across his substack and multiple reputable publications.

Executive takeaways

Artificial intelligence has become central to supply chain risk management. Companies are deploying predictive models, control towers, and agentic systems to monitor disruptions across suppliers, transportation lanes, ports, and global events. The infrastructure has never been more sophisticated. And yet a consistent complaint echoes across operations and planning teams: the AI flags the problem, but it cannot tell us what to do about it.

Building a context-ready architecture

Prabhat Rao Pinnaka is a product leader focused on building data and AI-driven enterprise platforms that enhance execution and decision-making across the end-to-end supply chain, including planning, procurement, warehousing, transportation, and customer fulfillment. He leads cross-functional teams in building analytics and AI-enabled workflow solutions that strengthen supply chain performance, increase operational visibility, and support governed automation at scale. Through his work as a keynote speaker, peer reviewer, and advisor, he contributes practitioner insight on the adoption of AI in operational environments. He serves on the ISCEA Americas Advisory Board and is a founding member of Saint Louis University’s Technology in Supply Chain Advisory Board.

Sukanya Bollineni is a Technical Product Owner at Johnson Controls. With a background in technology and product delivery, she works at the intersection of business needs, digital solutions, and cross-functional execution. Her professional focus aligns with broader efforts in digital manufacturing and connected industrial operations, where technology is increasingly used to improve efficiency, performance, and smarter decision-making.

Senthilkumar Thiyagarajan is a supply chain analytics professional whose work focuses on digital twins, supply chain optimization, and Industry 4.0 applications. Currently with Medline Industries, he brings a strong blend of academic research and practical industry perspective, with a Ph.D. in Supply Chain Management from Purdue University and a research focus on resilience in complex supply chains.

1. Ordibazar, A.H., et al. (2025). AI applications for SCRM considering interconnectivity, external events and transparency. Modern Supply Chain Research and Applications, 7(2), 148–179.
2. Kosasih, E.E., et al. (2024). Towards trustworthy AI for link prediction in supply chain knowledge graphs. International Journal of Production Research, 62(15), 2268–2290.
3. IBM Institute for Business Value. (2025). Alert Fatigue Reduction with AI Agents. IBM Think Insights.
4. Wu, J., et al. (2025). Enhancing supply chain visibility with generative AI: relationship prediction in knowledge graphs. International Journal of Production Research (online August 2025).
5. Nagy, J., et al. (2022). Increase supply chain resilience by applying early warning signals within big-data analysis. LogForum, 35(2), 467–481.
6. Aboutorab, H., et al. (2024). Text mining for proactive risk identification via NLP and reinforcement learning. Cited in Ordibazar et al. (2025).
7. Moktadir, M.A., et al. (2025). Systematic analysis of generative AI for supply chain transformation. Supply Chain Analytics, ScienceDirect.

Executive takeaways

Editor’s note: What It Really Means is a new recurring series on Supply Chain Management Review designed to clarify commonly used supply chain terms that often carry different meanings across organizations. The series aims to establish practical, shared definitions by grounding terminology in real-world planning and execution use cases. The series is authored by Andrew Byer, a former P&G supply chain leader, and Mike Dobslaw, who leads EY’s Supply Chain Planning Practice.

A phrase often cited in discussions to improve supply chain performance is “service is the essence of a supply chain.” But what does that really mean?

That service is the essence of a supply chain makes it clear the overarching purpose of a supply chain’s design and operations is to meet customer expectations. These expectations are translated into service targets. If service is not at target levels, the supply chain will be forced to adjust. These adjustments can include actions that can increase cost and cash needs to get service back to target. Cost would be impacted by things such as expediting freight, changing supplier and production schedules, running overtime, etc. Cash would be impacted by increasing safety inventories to buffer service. The net effect is without service being at target, it is hard for other key metrics to be at target. 

Why is the mindset that service is the essence of a supply chain important? 

A business will struggle if its customers cannot rely upon it to meet commitments to deliver products or services on specific timing and/or quantity. Unreliability is offset by the customer either risking their sales or carrying extra inventory. These offsets raise a potential double negative: (1) it creates an opening for competition who can better meet expectations and (2) it becomes harder for a ‘low service’ business to sell new products, services and innovation. Sales personnel know it can be very difficult to talk with a customer’s buyer about business-building ideas when the buyer is asking why you can’t reliably fulfill existing orders. In short, the business will struggle due to service issues. To the customer, the supply chain performance is indivisible from the overall company’s performance.

For this reason, service is usually a key metric on supply chain leadership scorecards. However, supply chain leadership juggles a lot of other functional metrics like cost, OEE, material unit price, inventory and new product speed to market—while trying to achieve a balanced scorecard.  

To this end, leading organizations differentiate target service levels across their products. These companies realize providing a high on-time/in-full-service target may not be achievable nor practical for their entire portfolio. It may come across as blasphemy, but having a differentiated service target approach that is aligned with customers, and measured and rewarded appropriately across functions is a step-change that many companies fail to achieve. This approach helps break the functional silos and behaviors sometimes seen in organizations that are rewarded and incentivized within their role: e.g. when manufacturing focuses on OEE/attainment instead of a critical changeover to meet a service goal.

From a priority perspective, however, providing customer service at target levels should supersede other metrics in overall importance. Note: leading-edge companies are able to align with their customers to translate how a customer measures service into their company’s internal target and tracking systems.

Benefits of the mindset that service is the essence of a supply chain: The supply chain is viewed as a competitive advantage and the enterprise is aligned cross-functionally around how to best meet service targets, creating tailwinds supporting business objectives, including:

• increased sales revenue and profit
• reduced expediting costs
• better customer partnerships and joint value creation opportunities
• easier to sell in new items and services
• clarity of purpose: overarching service metric is prioritized above functional metrics (functional groups could include manufacturing/operations, procurement, engineering, etc.)

Watchouts: Unfortunately, there can be many intended or unintended barriers to achieving a broad organizational mindset that service is the essence of a supply chain, including:

• Having a ‘one-size-fits-all’ approach to service goals. While service is the foundation, differentiated service goals should reflect product prioritization—with more important products having higher targets. All functions of the enterprise should be incentivized accordingly.
• Over-promising to customers vs. a supply chain’s demonstrated capabilities
• Not designing the supply chain (or insufficiently investing) to meet customer expectations e.g. instantaneous capacity, facilities close to market for responsive operations, etc.
• Commercial partners disengaged from what’s required to service customers (getting orders in on time, helping address master data or pricing issues that might block orders during processing, etc.)
• Drift on other key supply chain metrics. Just as a business will struggle if service is not in control, it also will have difficulty managing increased cost and cash needs on a going basis if required for service needs. 
• Supply chain sub-functional reward systems considering service “someone else’s job” (e.g. manufacturing seeking only long production runs, procurement only interested in lowest unit cost)

How to develop the mindset that service is the essence of a supply chain?

Consistently providing customer service that meets/exceeds customer expectations requires all facets of the supply chain to contribute.  Therefore, service needs to be the overarching goal and measure with no compromise if, for example, there’s a need to improve financial performance like cost or cash.  Functional and sub-functional reward systems all need to have service as a top-line measure.

About the authors

Andrew Byer is a former P&G Supply Chain Leader. Mike Dobslaw leads EY’s Supply Chain Planning Practice. To learn more about how EY and P&G team to support supply chain transformations please write michael.dobslaw@ey.com

Executive takeaways

Editor’s note: A full deep dive into this topic can be found on Knut Alicke’s substack. You can view that article here: https://knutalicke.substack.com/p/the-planner-was-the-system.

We have spent thirty years and tens of billions of dollars building advanced planning systems—constraint-based optimization solvers, multi-echelon inventory models, and statistical forecasting engines. In theory, these systems should produce better supply chain decisions than any human. They never get tired. They have no political relationships with the sales team that distort their judgment.

In practice, experienced planners routinely override the planning systems. And they are often right to do so.

The system produces a recommendation; the experienced planner looks at it, and something in their expression says, that’s not right. They change the number. The system logs the override, learns nothing, and produces the same wrong recommendation next month.

The usual explanations—poor data quality, insufficient user adoption, change management failures—are not wrong, but they are superficial excuses. The deeper answer is this: planning systems have never really understood the supply chain they were planning. They were given a structural skeleton and asked to make decisions that require a living, experience-rich understanding of how the supply chain actually behaves. So the experienced planner—carrying that understanding in their head—has been quietly doing the reasoning work the system was supposed to do. For 30 years.

Old planning systems ran MRP logic over relational tables. The supplier lead time of four weeks—one number, static. The system did not know that actual lead times ranged from two to eight weeks depending on the season and order volume. It generated hundreds of exception messages per planner per week and asked humans to do all the reasoning. The planner was the missing semantic layer.

The next generation modeled the supply chain as a knowledge graph—a genuine advance that enabled “why” questions the previous generation could never address. But the graph was still populated structurally. It knew the topology. It did not know that the Frankfurt office’s first-submission forecasts run 20% too high because the commercial team learned that supply chain applies a haircut and pre-compensates. The behavioral reality remained absent and required the planners’ experience and input.

The most recent generation adds probabilistic learning—genuinely valuable, but limited to patterns consistent enough to be statistically learnable. It learns that August lead times from a supplier in France are longer. It cannot understand why (summer vacation in France closes down everything). And for knowledge that lives in specific relationships and single painful experiences of the planner, there is simply not enough data for any model from which to learn.

Across all three generations, the behavioral knowledge lives in people’s heads, not in systems.

They have been with the company for 30 years. They know which suppliers can be pushed and which need careful handling. They’ve never written this down because writing it down would cause a political incident. They know a key customer always requests urgent delivery in the last week of the quarter to improve their own inventory numbers, costs that could be planned for if anyone paid attention to the quarterly calendar.

This is what Michael Polanyi called tacit knowledge: we know more than we can tell. Organizations have no systematic way to capture it, no mechanism to transfer it, and no plan for what happens when it retires.

This planner is probably retiring within five years.

Current planning systems have a structural ontology—they lack an experiential ontology: a knowledge layer that captures not what is formally true about the supply chain, but what has been learned to be behaviorally true through observation, experience, and organizational memory.

The structural ontology says supplier X has a lead time of four weeks, even modeling this probalistic would mean the average lead time is four weeks with a standard deviation of two weeks. The experiential ontology says supplier X has a behavioral lead time that runs to six weeks in Q4 due to factory capacity constraints during the harvest period. This pattern has been observed consistently for seven years. It is not in any contract. It is known only to planners who have managed this relationship over multiple Q4 cycles.

One more example—the most expensive kind. In week five of a six-week sea freight transit, inventory approaches the safety stock threshold. An inexperienced planner sees inventory falling toward what appears to be a danger zone and panics. They place an urgent air freight order at 10 to 12 times the sea freight cost. Four days later, both shipments arrive simultaneously. The safety stock was never breached. The air freight was never necessary. An experienced planner knows that safety stock exists precisely to cover demand volatility during an open replenishment lead time—approaching it while a confirmed order is in transit is normal behavior, not a crisis signal. A planner who does not understand this turns a routine week into an expensive one, every single time.

The experiential ontology does not replace the planning system. It completes it.

 

GenAI does not automatically build an experiential ontology. It is not magic. But it provides, for the first time, capabilities that make building one tractable: natural language conversation, reasoning over contextually rich information, and learning from interaction over time. Combined, they enable something specific: a system that can learn from experienced planners by talking to them.

Think of a good new associate joining the planning team. They do not walk in on day one and start making decisions. They observe. They ask questions. They earn trust gradually by demonstrating that their questions get smarter. Over the course of 18 months, they internalize an understanding that no training program could provide.

This is precisely how a GenAI layer should behave—not as a system that processes structured inputs and returns outputs, but as a genuine learning presence that observes planning decisions, asks questions at the right moments, and builds a progressively richer model of the supply chain’s behavioral reality. The crucial point is that planners cannot write their experience/tacit knowledge down. Ask a 30-year planner to document everything they know about a key supplier, and they will produce a paragraph. Engage them in a conversation about a specific late delivery in Q4, and they will produce an hour of structured, causally connected insight. The conversation is crucial

But the GenAI layer does not only learn from what actually happens. It can also create situations—drawing on the operational history encoded in the planning system, or constructing realistic disruption scenarios that routine operations rarely surface—and learn from how planners respond. When an experienced planner walks through their thinking about a force majeure from a sole-source supplier, or a sudden demand spike across three product lines simultaneously, they externalize a reasoning architecture that no behavioral observation would ever reveal. The system learns not just what they decided. It learns how they think. That is the difference between a passive recorder and a genuine new associate—and it is what makes the knowledge base genuinely compound over time.

In practice, this means capturing the reasoning behind every planner override—not through a form, but through a brief natural conversation. “You increased the safety stock for this material by 40%—is this a supplier concern, or something else?” The answer gets stored as a structured knowledge element with entity, behavioral pattern, condition, and—critically—implications for decisions downstream.

The design principle is non-negotiable: every interaction must feel like value exchange, not data extraction. The system must give back more than it asks for, immediately. A downstream risk identified from the planner’s reasoning. A flag that another planner handled a similar situation differently. The moment the interaction feels like filling in a form, engagement collapses.

The supply chains that build experiential ontologies now—in this period of volatility, disruption, and talent retirement—will have a compounding advantage that cannot be replicated quickly. Not because they deployed the best technology. Because they finally captured the best knowledge.

That knowledge is sitting in your planning team right now, in the head of someone who is probably not in any leadership presentation, who has probably been in the same role for 15 years, and who is probably thinking about retirement.

Before they leave, ask them why.

 About the author

Knut Alicke is a Senior Advisor to McKinsey, Partner Emeritus at McKinsey, a member of the Global Advisory Council for Alcott Globa, co-author of the book Source To Sold, and professor at the University of Cologne.

Executive takeaways

The conversation around artificial intelligence in supply chain and operations has shifted decisively. The question is no longer whether AI belongs in the enterprise; it's whether your organization can demonstrate measurable returns from it. And for most companies, the honest answer is: not yet. Despite record levels of investment, the gap between AI ambition and operational reality continues to widen.

Nathanael Powrie

Gartner reports that 72% of supply chain organizations have deployed generative AI, yet most are experiencing middling results for both productivity and ROI. Perhaps more telling: only 23% of supply chain leaders have a formal AI strategy in place. The rest are pursuing AI on a project-by-project basis, an approach that Gartner warns leads to fragmented architectures that hinder scalability and extend payback timelines.

The gap between AI spending and AI value creation is not a technology problem. It is an execution problem, and it is exactly the kind of problem that supply chain and operations leaders are positioned to solve.

The pilot trap: Why most AI use cases stall

Across industries, a common pattern has emerged. Companies launch AI pilots in forecasting, inventory optimization, or logistics planning. Early results look promising at the individual-task level. But scaling those pilots into enterprise-wide capability—the step that actually drives EBITDA impact—proves far harder than expected. The root cause is structural: most organizations are bolting AI onto legacy systems never designed for real-time decision-making.

Gartner's research confirms the disconnect: while GenAI tools save desk-based supply chain workers roughly four hours per week individually, those gains shrink to just 1.5 hours at the team level, with no correlation to improved output quality.

The lesson is clear. AI does not fail because the models are inadequate. It fails because the operational infrastructure—data governance, process standardization, and cross-functional alignment—has not been built to support it. This is a critical distinction for executives weighing their next round of AI investment. The technology is mature enough. The question is whether the organization is ready to absorb it.

The companies that stall at the pilot stage share a common profile: they invest in point solutions without first aligning on a unified data architecture or redefining the decision rights that AI is meant to enhance. The result is isolated pockets of automation that never connect to the broader planning and execution ecosystem.

Where AI is delivering real operational leverage

Despite the headwinds, companies that approach AI as an operational capability rather than a technology experiment are capturing significant value. The data points to three areas where returns are most tangible:

Demand forecasting and inventory optimization. Demand forecasting is the most mature AI use case in supply chain, with adoption rates reaching 87% among leading organizations. Companies embedding machine learning into their S&OP processes are seeing forecast accuracy improvements of 20% to 40%, translating directly into working capital release, reduced carrying costs, and improved service levels. Gartner predicts 70% of large organizations will adopt AI-based forecasting by 2030, but the leaders gaining the advantage are doing it now.

Procurement and supplier management. AI-powered spend analytics and supplier risk scoring are enabling procurement teams to identify savings opportunities that traditional category management misses. Top-performing supply chain organizations are investing in AI to optimize processes at more than twice the rate of low-performing peers, according to Gartner. These leaders use AI-driven decision-making to unlock new sources of value in supplier collaboration, category strategy, and contract management.

Real-time visibility and disruption response. AI-powered control towers are replacing static dashboards with predictive, self-correcting systems that autonomously reroute shipments or reallocate inventory the moment a disruption signal is detected. BCG reports that agentic AI systems accounted for 17% of total AI value in 2025, projected to reach 29% by 2028. However, Gartner notes that 23% of AI control tower projects stalled in 2025 due to a lack of cross-functional alignment, reinforcing that the technology works when the organizational foundation supports it.

The execution gap: What separates winners from the rest

The pattern among companies successfully scaling AI in supply chain operations is remarkably consistent. They do three things differently.

First, they standardize before they automate. Research consistently shows that the vast majority of AI initiatives struggle to deliver sustained ROI due to fragmented data, siloed systems, and undocumented workflows. Successful organizations invest in data governance and process harmonization before deploying AI, building the foundation that allows models to operate on clean, consistent inputs. Capgemini found that companies with a formal AI change management plan are 2.7 times more likely to achieve ROI within the first 12 months of deployment.

Second, they embed AI into existing workflows rather than creating parallel systems. The most effective deployments augment how planners, buyers, and operators already work. This is not about replacing people. It is about compressing cycle times, surfacing better options faster, and freeing experienced practitioners to focus on judgment-intensive decisions. Gartner predicts that 40% of enterprise applications will embed task-specific AI agents by 2026, up from less than 5% in 2025. But broad autonomous deployments without clear workflow integration are the ones that fail.

Third, they define success metrics before deployment and hold themselves accountable. Organizations that treat AI as a measurable investment, with defined cycle-time targets, documented cost savings, and CFO-trusted impact metrics, are the ones securing executive backing and scaling beyond pilots. Those that don't are seeing budgets reallocated. Gartner warns that 60% of supply chain digital adoption efforts will fail to deliver promised value by 2028, largely due to insufficient investment in change management.

2026: The year of accountability

The AI landscape is entering a new phase. The number of companies moving AI projects into production is accelerating, yet supply chain volatility shows no signs of abating. Everstream Analytics rates geopolitical fragmentation at a 97% threat level for 2026, while extreme weather risk sits at 93%.

This convergence creates a decisive moment for operations leaders. The companies that will emerge strongest are those treating AI as core operational infrastructure: embedded in planning, procurement, and logistics; governed with the same rigor as financial systems; and measured against hard P&L outcomes.

For procurement and supply chain leaders specifically, 2026 is the year that separates organizations that can demonstrate ROI from those that cannot. Executives who show faster cycle times, documented cost savings, and CFO-trusted impact metrics will secure continued investment. Those who cannot will find budgets redirected. The window for experimentation without accountability has closed.

The global AI-in-supply-chain market has grown from $6.5 billion in 2022 to nearly $20 billion in 2026, with projections exceeding $70 billion by 2030 according to Grand View Research. But investment alone does not create value. Companies that invest as much in people, processes, and change management as they do in technology consistently outperform those that lead with tools.

The question for every C-suite is whether their AI investment is translating into competitive advantage, or simply keeping pace with spend.

About the author

Nathanael Powrie is Senior Director of Knowledge Management and Data Analytics at Maine Pointe, a global supply chain and operations consulting firm. He leads AI-driven and data-centric initiatives that modernize supply chains by combining human expertise with intelligent automation, drawing on more than a decade of experience across manufacturing, logistics, and automotive industries.

Executive takeaways

In May 2024, FedEx announced a non-cash impairment charge of $157 million to permanently retire aircraft and engines. While part of a broader modernization, the charge underscored a brutal truth in logistics: assets are only valuable if they align with the network. When the network shifts, a sophisticated asset can become a liability overnight.

For decades, chief financial officers (CFOs) have evaluated supply chain automation such as automated storage and retrieval systems (ASRS), conveyor belts, and sorting hubs using a static playbook. The operations team presents a business case based on labor arbitrage and throughput; the finance team runs a discounted cash flow (DCF) analysis; and if the net present value (NPV) is positive, the capital is deployed. The underlying assumption is that the “logistics grid” of the roads, ports, and trade corridors outside the warehouse is a fixed constraint.

That assumption is no longer safe. We have entered the era of kinetic infrastructure.

From the $1.2 trillion PM Gati Shakti masterplan in India to the giga-projects of Saudi Arabia’s Vision 2030 and the reshoring incentives of the U.S. Infrastructure Investment and Jobs Act, governments are actively redrawing the map of global trade. They are not just paving roads; they are shifting trade gravity. A new rail corridor or a state-subsidized logistics park can render a legacy facility 50 miles away economically obsolete in months.

For the CFO, this introduces a silent but growing risk: the stranded logistics asset. This article argues that to prevent capital destruction, investment committees must move beyond static DCF models and adopt a “kinetic balance sheet” one that prices infrastructure risk and values flexibility as a premium asset.

The valuation trap: Duration mismatch

The root of the crisis lies in a “duration mismatch” between the asset’s useful life and the stability of the environment it serves.

Consider a standard automation project. A fully integrated ASRS system often requires a 10- to 15-year depreciation schedule to justify the massive upfront CapEx. However, in emerging markets and rapidly reshoring economies, the policy velocity—the rate at which infrastructure and regulations change—is effectively 3 to 5 years.

When a CFO approves a 15-year fixed asset in a region with a 5-year policy horizon, they are essentially “shorting” volatility. If the state commissions a new digital customs platform that your legacy system cannot plug into, or opens a freight corridor that bypasses your facility, your location premium evaporates. The asset still works mechanically, but it is strategically dead. It has become a “fixed-cost albatross” unable to compete with rivals in the new corridor, yet too expensive to abandon.

The failure of static DCF

Standard DCF models are ill-equipped to handle this volatility for three reasons:

1. Linearity bias: DCF assumes a steady state of operations. It cannot easily model binary, structural breaks like a government suddenly banning a specific class of truck or mandating a new digital interface.
2. Zero value on flexibility: In a DCF spreadsheet, a $10 million investment in rigid, bolted-down conveyors looks identical to a $10 million investment in flexible, autonomous mobile robots (AMRs). In reality, the risk profiles are polar opposites. The conveyor is a sunk cost; the robots are liquid assets that can be moved. DCF assigns zero financial value to this ‘option to switch.’
3. Uniform discount rates: Most investment committees apply a firm-wide weighted average cost of capital (WACC) to all supply chain projects. This ignores the fact that a warehouse in a stable region (e.g., Hamburg) has a radically different risk profile than one in a high-flux zone (e.g., Uttar Pradesh or Riyadh).

To fix this, CFOs need to introduce new metrics into the investment committee lexicon.

New metric 1: Infrastructure beta ( )

In portfolio theory, beta measures a stock’s volatility relative to the market. In logistics, we need infrastructure beta ( ): a measure of an asset’s sensitivity to changes in the public logistics grid.

• High-beta assets: These are facilities whose value is tightly coupled to a specific piece of public infrastructure.
  • Example: A cross-dock facility built exclusively to service a specific rail siding. If that rail line becomes congested or the government changes the tariff regime for rail freight, the facility’s value collapses. It has no redundancy.
  • Financial implication: High-beta projects carry higher systemic risk. They should be assigned a higher hurdle rate (discount rate) in the valuation model to compensate for the lack of resilience.
• Low-beta assets: These are assets that are decoupled from specific infrastructure shocks.
  • Example: A “dark store” for rapid delivery in a dense urban center, utilizing a fleet of electric bikes. Its value is derived from proximity to population, not a specific highway. If one road closes, the bikes take another.
  • Financial implication: These assets are defensive. They deserve a lower hurdle rate because they preserve cash flow even during infrastructure disruption.

The CFO’s move: Audit your current CapEx portfolio. If your balance sheet is loaded with high-beta assets in high-velocity markets, you are over-exposed to sovereign execution risk.

New metric 2: The asset flexibility premium

If volatility is the enemy of fixed assets, flexibility is the hedge. This brings us to the financial structure of automation.

Historically, finance teams have preferred CapEx for automation because owning the asset allows for depreciation and EBITDA normalization. Conversely, Robotics-as-a-Service (RaaS), a subscription model where you pay for robots as you use them, is often viewed skeptically because it hits operating expenses (OpEx) and can appear more expensive on a total cost of ownership (TCO) basis over 7 years.

This TCO view is flawed because it ignores the value of optionality.

In financial terms, a RaaS contract is not just a rental agreement; it is a bundle of put options. A put option gives the holder the right to sell an asset at a specific price. When you sign a RaaS contract with a 6-month cancellation clause, you effectively own a put option that allows you to “sell” the asset back to the vendor at zero cost if the market turns against you.

Consider two scenarios for a company building a hub in a developing logistics zone where the government might shift the primary freight corridor:

1. The CapEx owner: You build a fixed system. Three years later, the corridor moves. You are stuck with a depreciating building and steel racking that costs more to dismantle than it is worth. You face a massive write-down.
2. The RaaS subscriber: You exercise your put option. You cancel the subscription, return the robots, and move your operations to the new corridor with minimal capital destruction.

The CFO’s move: Calculate the “strategic NPV” of RaaS.

In high-volatility environments, the premium you pay for RaaS is simply the cost of the insurance policy against obsolescence. It is almost always cheaper than a write-down.

The investment committee checklist

To operationalize this, the investment committee must change the questions it asks project sponsors. Stop asking, “What is the ROI?” and start asking, “What is the liquidity of this asset?”

Here is the new 4-point diagnostic for capital approval:

1. The “off the map” test
   • Question: Is this site located on a corridor prioritized in the national logistics masterplan (e.g., Gati Shakti layers)?
   • Red flag: If the site is “off the map” and relying on legacy roads while the state builds highways elsewhere, reject the CapEx. The facility will face structurally higher logistics costs than competitors aligned with the new grid.
2. The interoperability stress test
   • Question: Does this automation system have open APIs to plug into future public data platforms (e.g., single window logistics portals)?
   • Red flag: “Black box” proprietary systems. If the asset cannot ingest real-time public data (port berthing times, rail visibility), it will suffer from information asymmetry. It is a “data blind” asset.
3. The liquidity audit
   • Question: If we lose our anchor customer in Year 3, what is the secondary market value of this equipment?
   • Red flag: Custom-bolted steel infrastructure. If the cost to dismantle and move the equipment exceeds 50% of its value, it is illiquid. Prioritize modular, mobile units (AMRs) that preserve balance sheet agility.
4. The contract-asset match
   • Question: Does the asset’s depreciation schedule match the duration of the client contract?
   • Red flag: A 10-year asset supporting a 3-year contract. This duration mismatch is the primary cause of impairment charges. Push for RaaS to match expenses directly to revenue.

Conclusion: From gatekeeper to architect

The supply chain of the future will not be defined by who has the fastest robots, but by who has the most agile balance sheet. We are witnessing a decoupling of “production” from “location.” As governments rewire the physical world, the most dangerous risk a company can take is to pour concrete into a map that is being redrawn.

For the CFO, this is an opportunity to move beyond the role of gatekeeper. By adopting metrics like infrastructure beta and valuing the optionality of flexible automation, finance leaders can stop their companies from investing in the past and start building a portfolio resilient enough for the kinetic future. Location is no longer just a place; it is a financial derivative. Value it accordingly.

The “infrastructure beta” of your portfolio

Use this logic to classify your assets during the next quarterly review.

Executive takeaways

The Organization for Economic Co‑operation and Development (OECD) predicts that artificial intelligence could increase global exports by as much as 40% by 2040. As AI tools are adopted across industries and geographies, they are reshaping how business and government leaders approach trade, economic growth, and national security.

At the same time, a new class of market access barriers is emerging. Unlike traditional barriers such as tariffs or physical borders, these obstacles stem from regulatory fragmentation, unequal access to digital infrastructure, and uneven availability of capital and advanced computing resources. AI adoption in advanced economies is expanding nearly twice as fast as in developing countries. As a result, countries and firms capable of building and deploying AI systems are pulling further ahead of those that cannot.

This growing divide is creating structural challenges for global trade and digital supply chains. Governments and supply chain leaders must now navigate a world in which competitiveness increasingly depends on access to data, computing power, and AI capabilities.

The AI divide and the cost of keeping up

As artificial intelligence adoption accelerates, an AI divide is emerging between organizations and countries that can deploy these technologies and those that cannot. This divide is creating a new set of operational and strategic challenges for supply chain and government leaders worldwide.

Building and running AI systems requires substantial capital investment. Data centers depend on components whose costs have risen roughly 40% over the past five years. Access to electrical grid connections is also becoming increasingly constrained. In some regions, connection waitlists now extend up to seven years.

Several countries—including Singapore, Ireland, China, the Netherlands, and Canada—have introduced restrictions on data center development, location, or energy use in response to concerns about power demand and environmental impact. At the same time, electricity demand from AI computing infrastructure is projected to grow dramatically over the next decade.

Advanced economies and large enterprises are far better positioned to absorb these costs. Many emerging economies and smaller firms lack the infrastructure, capital, and technical talent required to deploy AI systems at scale. According to the OECD, approximately 66% of firms in high-income economies report using AI, compared with 27% in lower-income economies.

This widening capability gap will increasingly shape supply chain performance and global trade flows. Regions with strong AI capabilities are likely to see greater efficiency, faster innovation, and improved resilience. In contrast, regions that lag in digital infrastructure may experience slower productivity gains and persistent operational bottlenecks.

Dive deeper

Dravida Seetharam and Sarah Lahti join the Talking Supply Chain podcast to talk about AI-driven trade barriers.

For governments, the implications are equally significant. Countries that fail to invest in digital infrastructure and AI capacity risk slower economic growth and declining competitiveness in global markets.

Cybersecurity risks

The rapid integration of AI into supply chains is also increasing cybersecurity risks.

The World Economic Forum’s Global Cybersecurity Outlook report found that 87% of respondents identified AI-related vulnerabilities as the fastest-growing cyber risk. In addition, 64% of organizations now account for geopolitically motivated cyberattacks in their risk mitigation strategies.

As AI becomes embedded in logistics platforms, digital trade systems, and cross-border data exchanges, the attack surface for cybercriminals expands. AI tools can also be exploited by attackers to conduct more targeted and automated attacks.

For supply chain leaders, this means cybersecurity strategies must evolve rapidly. Risk management frameworks, security monitoring systems, and governance policies must continuously adapt to an operating environment where both threats and technologies are advancing quickly.

How governments are responding

Governments around the world are still determining how best to manage the economic and security implications of AI.

According to the Global Cybersecurity Outlook survey, 31% of respondents reported lacking confidence in their country’s ability to respond to a major cyber incident, up from 26% the previous year. This uncertainty has prompted many policymakers to explore stronger digital sovereignty measures.

Europe, for example, has launched EuroStack, an initiative aimed at reducing dependence on foreign technology providers and strengthening regional digital infrastructure.

While such policies may improve security and technological independence, they also risk fragmenting the global digital economy. Differing regulatory frameworks for AI systems, data governance, and cloud infrastructure can create additional compliance burdens for companies operating internationally.

A joint survey conducted by the World Trade Organization and the International Chamber of Commerce found that nearly one-fifth of surveyed firms identified regulatory fragmentation and uncertainty as major barriers to adopting AI technologies.

If regions pursuing digital autonomy cannot match the resources currently provided by global technology ecosystems, they may inadvertently restrict access to the large datasets and computing scale required to fully realize AI’s economic benefits.

Steps to mitigate AI-driven market access barriers

While many AI-related barriers are driven by national policy and infrastructure constraints, supply chain leaders can take several proactive steps to protect their organizations’ market access and operational resilience.

The strength of a supply chain depends on the capabilities of every partner within it. Organizations should assess the digital readiness of key suppliers, logistics providers, and markets to identify vulnerabilities before they become operational failures. Equal attention should be given to strengthening internal capabilities by ensuring supply chain professionals are equipped to work effectively with AI tools, data systems, and digital risk frameworks.

Companies should also diversify their digital infrastructure partnerships (for example, by adopting regional cloud strategies) to reduce exposure to regulatory or geopolitical disruptions. Organizations that address these priorities with urgency will be better positioned to manage disruptions and sustain a competitive advantage in the years ahead.

Conclusion

Artificial intelligence is reshaping the foundations of global trade faster than many organizations can adapt. Rising infrastructure costs, the widening AI divide, growing cybersecurity risks, and increasingly fragmented regulatory environments are all creating new challenges for supply chain leaders.

There is no single solution that addresses all of these issues. However, leaders who recognize these emerging barriers and proactively adjust their strategies will be better positioned to maintain competitiveness and resilience.

As AI continues to transform the global economy, the distinction between countries and companies that help shape the rules and those that simply follow them may prove as important as the technology itself.

About the authors

Dravida Seetharam, is a fellow at the Center for Global Enterprise.

Sarah Lahti is the Director of Operations and Program Management for the Digital Supply Chain Institute.

Executive takeaways

For years, procurement progress has been measured by how much of the function has been digitized. Over the last 30 years, a large proportion of core processes have moved into enterprise systems. Data is more connected, information is more accessible, and execution faster than it was even a decade ago. But artificial intelligence is changing the definition of progress.

If we step back for a moment, an uncomfortable truth emerges: the fundamentals of procurement haven’t really changed.

Requisitions still become purchase orders. Purchase orders still move through approval chains. Suppliers still invoice. Finance still performs a three-way match before payment. The technology has improved, but the underlying process logic remains largely intact.

Recent Economist Impact research shows that 68% of C-suite leaders rank AI proficiency and ethics among their top development priorities for the next 12 to 18 months, while geopolitical instability remains the most immediate risk focus for procurement leaders.

Procurement finds itself operating at the intersection of resilience, efficiency, and risk—often with fewer resources than before.

Automate where it makes sense

The fastest and most tangible value from AI will come from disciplined automation, not from trying to reinvent procurement overnight.

Digitally mature areas such as procure-to-pay (P2P) and sourcing are obvious starting points. These processes are transactional, rules-based, and data-rich. AI can help streamline invoice matching, guide buying behavior, flag anomalies, and automate elements of supplier selection.

Economist Impact data already shows progress. Over the past 12 to 18 months, organizations reported AI-driven improvements in productivity and efficiency (75%), cost optimization (75%), and contract management (75%), with particularly strong performance in source-to-contract automation (67%). Recent research from Kearney helps explain why these gains are concentrated in execution: most organizations are still applying AI to existing workflows rather than redesigning procurement as an end-to-end system, limiting impact to efficiency rather than structural advantage.

These gains matter, but they are, by definition, execution-level gains. They make procurement faster and more efficient. They do not, on their own, change the strategic posture of the function. That opportunity sits upstream.

Why strategy still belongs to humans

The real value in procurement doesn’t come from running events more quickly. It comes from deciding which events are worth running in the first place.

Category strategy—not tactical sourcing—is where the hardest decisions sit. In many markets, the balance of power increasingly favors suppliers, and traditional competitive sourcing approaches often underperform. Economist Impact research shows organizations are placing greater emphasis on resilience alongside cost, with geopolitical exposure more than doubling year-over-year as a top risk concern.

But it’s not like cost pressure has gone away. Cost savings were identified as the primary value proposition of procurement according to the Economist Impact report and remain a primary driver of external workforce and services procurement decisions. This creates tension and trade-offs.

In a constrained supply market with limited competition, defaulting to a competitive event can actually weaken your position. A more resilient approach might involve supplier development, nearshoring, dual sourcing—or, in some cases, revisiting insourcing decisions to regain control over critical capabilities. These are not sourcing decisions. They are business decisions.

This is where AI plays a different role. AI can synthesize market intelligence, model supplier concentration risk, simulate demand shocks, and run long-term cost and risk scenarios. It can help identify which subcategories justify diversification and which are better served through deeper partnerships. But it doesn’t replace judgment. It sharpens it. That distinction—automation vs. augmentation—will shape what procurement looks like next.

AI doesn’t fix a broken blueprint

For years, procurement has digitized existing processes without fundamentally redesigning them. AI makes that incremental approach harder to sustain.

If execution can be automated, what does that mean for team structure? If category strategies move from broad groupings to more granular sub- and micro-segments, how should work be organized? How do you maintain visibility and governance as supplier ecosystems expand without simply adding headcount?

These questions are already playing out in many organizations. Research indicates procurement headcount has remained flat or declined even as complexity increases. At the same time, reliance on services procurement, external workforce models, and broader supplier networks continues to grow.

The equation most procurement teams face is uncomfortable but familiar: more suppliers, more services, more risks—and fewer internal resources. Without changes to how work is structured, that equation becomes difficult to sustain.

Rather than layering AI onto existing models, organizations need to rethink how work is designed. That means clearer separation between strategic and transactional activity, tighter alignment between category strategy and enterprise risk planning, and more disciplined management of services procurement as outsourcing expands.

Turning mandate into momentum

The biggest risk heading into 2026 isn’t technological failure; it’s loss of momentum.

Most executives agree that AI adoption is necessary. Fewer can clearly articulate how it will be implemented at scale. Economist Impact research shows that confidence in procurement’s category management capabilities declined year over year, reflecting the growing complexity of the environment procurement operates in.

For procurement leaders, this is where the real work sits. Moving beyond pilots. Investing in data foundations before layering on AI capabilities. And developing teams that combine digital fluency with commercial judgment because one without the other doesn’t deliver outcomes.

AI won’t do the thinking for procurement. But it has a way of exposing where thinking has, or hasn’t, already been done.

The leadership test of 2026

Procurement has always balanced cost and resilience, shifting emphasis as conditions change. What’s different now is speed. Geopolitical volatility, supply fragmentation, and digital acceleration are compressing decision cycles.

In this environment, leadership won’t be defined by how many AI tools are deployed. The procurement leaders who stand out in 2026 will be those who automate what is repeatable, free up capacity where judgement matters, and are willing to challenge processes that haven’t fundamentally changed.

AI is not a bolt-on efficiency tool. It is a catalyst for structural change. And procurement’s next ambition isn’t about doing the same work faster—it’s about deciding which work still makes sense to do at all.

About the author

Gordon Donovan is Vice President of Research for Procurement and External Workforce for SAP. An accomplished leader with over 30 years’ experience optimizing procurement and supply chain operations and strategy, he has a well-established reputation for driving transformational change and building capacity across global organizations.

Executive takeaways

The NextGen Supply Chain Conference has named Tractor Supply Company as the recipient of its 2026 Visionary Award, recognizing the company’s supply chain strategy and operational initiatives supporting long-term growth.

The Visionary Award is the conference’s highest honor and recognizes organizations demonstrating leadership across supply chain strategy, technology, and organizational development.

Tractor Supply will accept the award at the NextGen Supply Chain Conference, Oct. 21–23, 2026, in Nashville, Tennessee.

Accepting on behalf of Tractor Supply will be Colin Yankee, EVP of Supply Chain. Yankee will participate in NextGen’s Visionary Keynote fireside chat and will discuss Tractor Supply’s supply chain journey.

Registration for this year’s event is ongoing. Click here to register.

The NextGen Supply Chain Conference is a practitioner-driven event designed for senior supply chain, operations, procurement, and logistics leaders attended by 250 senior supply chain executives, solution providers, consultants and academics from some of the nation’s leading supply chain organizations, companies and universities.  

The 2026 event will focus on four industry verticals: Logistics and fulfillment, retail, food & beverage and chemical/pharmaceuticals. Speakers will focus on real-world execution, implementation insights, and measurable outcomes.

In additional to Colin Yankee, early speakers announced to date include Carey Boone, VP-Transformation for the Americas at DP World, Andy Moses, SVP of Sales and Solutions for Penske Logistics, Bijoy Sasidharan, Director of Capacity Planning & Forecasting at Fanatics, Jay Di Sieno, Senior Supply Chain Manager at Berry Direct (Edible Arrangements), Norman Katz, Founder, President & CEO of Katzscan Consulting, and Dan Pellathy, researcher at the University of Tennessee’s Global Supply Chain Institute.

More speakers will be announced soon. If you are a senior supply chain professional working in one of the four focus areas and are interested in speaking, you can submit a proposal here.

How to participate

Organizations interested in participating in NextGen can:

• Register to attend
• Apply to speak
• Explore sponsorship opportunities
• Submit nominations for open award categories (deadline: May 15)

Tractor Supply aligns with business growth

Tractor Supply has outlined a multi-year strategy, “Life Out Here 2030,” which includes continued investment in supply chain capabilities to support store growth, direct fulfillment, and customer experience. As part of that strategy, the company has emphasized the role of the supply chain in supporting broader business objectives, including network expansion and service improvements.

The company continues to invest in its distribution network, including new and expanded facilities designed to support increasing demand and operational scale. These investments are intended to improve efficiency, capacity, and service levels across its growing store base and customer footprint.

Developing last-mile delivery capabilities

Tractor Supply has also expanded its last-mile delivery network, including the growth of hub locations that support delivery operations. These capabilities are designed to improve delivery coverage and service, particularly for larger and heavier products and for customers in rural and geographically dispersed areas.

To support this expansion, the company has invested in distribution facilities and operational capabilities, including the use of automation and other technologies to support efficiency and scalability within its network. These efforts reflect a continued focus on improving operational performance while supporting long-term growth.

Recognizing supply chain leadership

The Visionary Award is part of a broader NextGen awards program that recognizes excellence across multiple areas of supply chain execution, including:

• Intelligent Transformation (AI and digital technologies)
• Autonomous Operations (robotics and automation)
• Startup innovation
• Partnership in Execution, a new category for 2026

Entries are now being accepted for these additional awards. Interested companies can apply here. More information, including descriptions of the awards and eligibility and evaluation criteria, is available here.

Together, these awards highlight organizations delivering measurable results and advancing supply chain capabilities across the industry.

More information is available at nextgensupplychainconference.com.

Executive takeaways

Editor’s note: This article first appeared on the University of Tennessee, Knoxville’s Global Supply Chain Institute’s blog. It is being reprinted with permission. You can read the original post here.

Authors' note: In September, the University of Tennessee Global Supply Chain Institute published “The 5Ps of Returns Management: An Integrative Approach,” by Huseyn Abdulla and Tom Goldsby. The white paper summarizes research conducted by the Advanced Supply Chain Collaborative, a think tank that partners faculty experts with industry professionals to enhance business performance, enrich teaching, and cultivate supply chain excellence.

Returns are a fact of life in retail and e-commerce. Nearly 17% of online purchases are returned, costing companies billions while straining customer relationships. But what if returns could become a source of value rather than a headache?

A new white paper from the UT Global Supply Chain Institute shows that leaders can reduce costs and improve customer loyalty by managing returns holistically. Authors Huseyn Abdulla and Thomas J. Goldsby introduce the 5Ps of Returns Management framework—people, policies, processes, products, and partners—as a roadmap for turning reverse logistics into competitive advantage.

Here’s what supply chain professionals should consider when tackling returns:

• People. Train frontline associates and customer service teams to make fair, consistent return decisions. Human judgment is the starting point for every return
• Policies. Strike the right balance between leniency and protection. Overly strict policies may cut return rates but drive customers away
• Processes. Don’t just process returns—design them. From returnless refunds to in-store drop-offs, smart processes reduce friction and capture value
• Products. Customize policies by category. A one-size-fits-all return window doesn’t reflect differences between apparel, appliances, or seasonal goods
• Partners. Work with third-party providers to detect and prevent return abuse as well as for refurbishment, donation, and resale of the returned goods

Their research, produced for the Advanced Supply Chain Collaborative, highlights that managing any one of these elements in isolation is rarely enough. Instead, it’s the interaction of all five together that determines whether a company will succeed or struggle with returns.

Consider Amazon’s partnership with Kohl’s. Designed to provide hassle-free returns for Amazon customers, the arrangement initially boosted store traffic for Kohl’s. However, it also created long lines, strained employees, and didn’t deliver the profits Kohl’s expected: an example of how one dimension of returns management, if not carefully balanced with others, can backfire.

By contrast, Philips Consumer Electronics in the early 2000s faced massive returns problems but solved them by implementing a coordinated, cross-functional strategy. The company created a dedicated returns management team, standardized policies, partnered with retailers, and integrated consumer feedback into product design. The result: a $100 million annual reduction in returns.

“The best return is the one that never happens,” Abdulla and Goldsby note. But when returns do happen, using the 5Ps ensures they’re managed strategically rather than reactively.

Returns management doesn’t have to be a drain on your supply chain. With the right policies, processes, and partnerships in place—and a strong focus on people and products—companies can transform returns from a cost center into a source of value.

Download the white paper, “The 5Ps of Returns Management: An Integrative Approach,” to explore case studies and practical strategies you can apply today.

To learn more about how your company can partner with the Advanced Supply Chain Collaborative to explore advanced concepts in supply chain management, visit ASCC.

Executive takeaways

Europe is in reindustrialisation mode, with new EU-wide initiatives such as the Clean Industrial Deal and the Industrial Accelerator Act aimed at rebuilding strategic capacity and competitiveness. But while much of the policy debate is focused on energy costs and regulation, another vital factor requires attention: the reliable movement of goods within and through the continent.

Rashid Abdulla

European Commission president Ursula von der Leyen, in her address at this year’s World Economic Forum, pointed out that while capital and data can cross the continent in a second, business expansion is complicated by “a new set of rules every time” a company enters another EU Member State. This acts as “a handbrake” on growth and profit potential.

This problem with “a new set of rules every time” also applies to supply chain networks in Europe. When documentation, checks, capacity access and compliance differ across borders, lead times become harder to predict. Faced with that uncertainty, companies self-insure themselves: they carry more buffer stock and tie up more cash in working capital so factories and customers aren’t left waiting when the network stutters.

Rail freight: an underused strategic asset

Rail freight illustrates this friction. With more than 200,000km of track in the European Union alone, Europe’s rail network should be its strategic strength; as the natural connector between ports and inland industrial regions, it is the obvious efficient, more sustainable alternative to congested road and water networks.

Yet for many cargo owners, cross-border rail is more challenging than it should be. Freight trains will run smoothly within one country and then lose hours during the planned transfer time at the border. It is no surprise that rail’s share of EU freight fell from 18.9% to 16.9% between 2018 and 2023, with only six countries increasing the modal share of rail freight between 2005 and 2023. The consequence of this is that reliability—a supply chain’s most valuable currency—suffers.

Turning plans into performance

Still, there is reason to hope we are at a turning point. The EU’s Trans-European Transport Network (TEN-T) sets out a continent-wide plan for transport corridors that link countries and modes, while the Connecting Europe Facility (CEF) provides the funding to build and upgrade the cross-border projects that make those corridors work in practice.

For example, the Brenner Base Tunnel targets a major north-south bottleneck between Austria and Italy and is designed to make cross-border rail slots predictable enough to compete with road, by replacing a steep, winding 19th-century mountain alignment with a nearly flat, straight base tunnel—increasing speed and load capacity, shortening the route, and cutting transit times. Rail Baltica does something similar in the north-east by building a route that connects Baltic ports and inland markets across Estonia, Latvia, Lithuania and Poland to the wider EU network.

At DP World, we are committed to this next generation of trade corridors. For example, in Romania, the recent expansion of the Port of Constanţa alongside our Aiud inland hub provides a direct, high-tech link between the Black Sea and Central Europe’s industrial centres. While at Southampton in the UK, our Modal Shift programme has removed 137,000 truck journeys from the road in little over a year using the UK’s underutilised rail network to connect businesses to customers faster.

However, a world-class tunnel or port is only as reliable and efficient as the digital systems governing it. These systems must be interoperable, which is what the Electronic Freight Transport Information (eFTI) framework seeks to achieve. By replacing paper-based friction with a unified digital exchange, the eFTI could save the transport sector an estimated €1 billion annually in admin and compliance costs. This is promising, yet mandatory acceptance of digital freight data will not apply across all EU Member States until mid-2027.

Data, decarbonisation and delivery

One further reason to get the physical and digital layers right now is the EU’s carbon policy. Under the expanded Emissions Trading System 2 (ETS2), BloombergNEF projects average EU road-transport fuel prices could rise by almost one-third by 2030, potentially adding €0.50 per litre of fuel. In that environment, a reliability gap also quickly becomes a cost gap, and the case for switching to more sustainable multimodal routes strengthens.

So we have the diagnosis. We have signs of action. But there’s a lot more to be done. Deutsche Bank analysis shows that of the 900 measures recommended by the landmark 2024 Draghi report to fix our supply chains, only 11% are in place.

To restart Europe's industrial engine, the public sector must take the lead: standardise rail to work across borders; treat data as a utility by fast-tracking digital freight standards (eFTI); and fund more corridors that move goods. But the private sector cannot wait on policy alone. Whether you’re a cargo owner or port operator, now is the time to invest in getting the digital and physical layers future-ready.

Europe’s competitiveness depends on the ability to deliver on its extraordinary growth potential. The assets are there, we now need to ensure that the connection is there as well. We might not be able to move goods as fast as capital and data, but there’s no reason why we shouldn’t aspire to have them follow closely behind.

About the author

Rashid Abdulla is CEO and MD of DP World Europe where he oversees a pan-European network spanning ports and terminals, contract logistics and freight services. He previously led the Asia Pacific region and served as SVP Global Operations at DP World’s Head Office.

Executive takeaways

The closure of the Strait of Hormuz has roiled global energy markets. Diesel fuel is among the most affected, with prices increasing 46% from the first week in February. Though there has been much alarm over the likely economic effect, the impact of higher diesel prices on motor carriers and supply chains will be minimal. Sudden prices swings are disruptive, but U.S. transportation networks can sustain higher diesel prices.

The first question to address is who ultimately pays the higher cost of diesel in trucking markets. The answer is shippers. Whether it is contract markets, where diesel is priced explicitly and separately, or spot markets, where it is not, diesel costs are quickly and entirely passed through. There are three important pieces of evidence for this. 

First, we can look at the data on how temporal variation in diesel prices translates to variation in spot market prices. This is shown in figure 1, where diesel prices are translated into dollars per mile, assuming a factor of 6 miles per gallon. Statistical analysis shows that 50% of the change in diesel prices is immediately passed through.  Within a week, the pass-through rate hits 100%.

 

The second piece of evidence on pass through is based on state-level variation in diesel taxes. California has a high tax rate on diesel. Mississippi has one of the lowest diesel tax rates. Other things equal, you would expect truckload rates in California to be higher than those in Mississippi. Research has found exactly this; moreover, the pass-through for these taxes is complete—even over-shifted so that shippers pay a markup.

Third, there is a theoretically compelling argument for full pass-through. The motor carrier industry is extremely competitive, with thousands of motor carriers entering and exiting the market each year. This competition makes profit margins so razor-thin that industry supply is dictated by the cost of service. Shipper demand for freight transportation, on the other hand, is comparatively less sensitive to price. Based on economic theory alone, we expect full pass through. Motor carriers are not in a position to absorb the diesel costs because they would suffer losses and go out of business. While some cargoes may go unshipped during periods of high prices, most freight still must move.

Motor carriers do suffer losses and go out of business all the time. Volatile prices are difficult to manage, especially for smaller motor carriers that do not have dedicated staff of market researchers and price analysts. It takes time for them to adjust to the new reality. Fleet fuel efficiency also matters. High mpg fleets can undercut the less fuel-efficient fleets, so reduce your speed and deploy those boat tails!

These three pieces of evidence paint a compelling picture of an industry that adjusts to fluctuations in fuel prices. While that is reassuring for motor carriers, it does leave a question of the aggregated effects.

Will these cost increases show up in higher consumer prices? You won’t notice it. On average, transportation (all of it) makes up 3-4% of the value added in the country. The items where transportation does account for a significant share of costs (fresh fruits and vegetables, say) make up a small share of household budgets.

Energy prices have long been volatile. Transport markets and supply chains have evolved to be robust to the inevitable price swings.

About authors

Andrew Balthrop is an assistant professor of supply chain management at the University of Tennessee, Knoxville, where he teaches logistics operations management. His research focuses on safety and environmental policy in the trucking industry.

Timothy Fitzgerald is an associate professor of economics at Baker School of Public Policy and Public Affairs at the University of Tennessee, Knoxville where he teaches energy economics.  He served as a senior economist on the White House Council of Economic Advisors, 2017-2018.  His research focuses on natural resource and energy economics.

Executive takeaways

Pickup defects, which we define at late pickups from a vendor or shipper site, represent a very costly challenge in transportation cost management. Pickup defects often cascade to late departures from shipper or vendor sites and late arrivals at downstream sites, and can lead to customer promise misses.

They impact vendor experience and additional volume from suppliers. Pickup delays can also cause labor and volume availability mismatches at downstream sites, inventory placement issues and also slow down inbound transportation speed. U.S. logistics costs reached $2.58 trillion in 2024 (about 8% of GDP) and transportation costs reached $937 billion (CSCMP State of Logistics report, 2025). Even a small percentage improvement in pickup defects can reduce overall transportation cost for the industry by manifold.

This article introduces a new framework called PRI (Predictive Reliability Index) that leveraged machine learning to predict carrier-level risk in future pickup defects. By analyzing more than 147 input factors that affect pickup defects across 150,000-plus records in various logistics network, the model attempts to enable proactive actions to prevent pickup defects from occurring. While 100% defect mitigation is not possible from the model’s output yet, pilot results showed a 35% reduction in pickup defects among high-risk carriers. With an accuracy rate of 85%, the model is projected to deliver cost savings of $40M-plus at full-scale implementation. Until PRI, most networks across logistics companies focused on on-time delivery used only reactive means to resolve issues and take actions to prevent defects.

This paper outlines the problem landscape, the analytical methodology, and recommendations for supply chain practitioners considering similar predictive approaches.

Why reactive carrier management does not work

Carrier performance in most logistics networks is reactive in nature. Drivers arrive late to their destination, the defect is identified either through escalations from the field or operations teams, or through periodic data analysis. Corrective actions are identified and implemented, but by that time, using a recovery load might be too late to protect a customer promise. Also, if a load arrives late to its destination, the operations teams might already be using available labor capacity to process loads that were not a priority. So, even when the load arrives, the operations team might not be able to process the load immediately, leading to further delays. These issues can also create inventory placement issues or yard utilization risks if the load is dwelling in the yard. In the pilot study, we estimated that every unit that dwelled for an hour in the yard cost between 29 cents to 45 cents. That’s why proactively preventing defects is the most cost-effective solution.

There has been some research in the industry supporting this—in a 2023 analysis performed by CSCMP, the researchers found that one of the top three causes of inventory placement and availability issues in retail and manufacturing supply chains was inbound transportation delay. Also, research published in the Journal of Business Logistics has suggested that carrier performance and on-time arrival are predictable based on historical data.

Why do big logistics and retail companies continue to be reactive in that case? The primary reason is the unavailability of a singular source of data to perform a holistic evaluation of these defects. Different teams within a company track different metrics, and they usually take ownership of their part of the problem only. This leads to different teams creating different root cause analyses for the same defect. For example, pickup defects can be related to vendor sites not providing on-time check-in to the drivers, the scheduling team creating an incorrect schedule, the carrier team dispatching a driver late or a driver arriving late due to weather or other issues. Each team only addresses its part of the problem. Also, each team’s manager relies on dashboards specifically built for their team to reactively find solutions.

We analyzed a large, very complex inbound network for the study. The analysis found that 1.35% of carriers were responsible for a large percentage of the pickup defects, based on data from over 1,600-plus carriers. Without a machine learning model, it was difficult to identify these high-risk carriers since new carriers are frequently onboarded and many carriers experience pickup defects for the first time each week. This finding is consistent with what current academic research in supply chain literature points to.

Development of the PRI framework

Development of the PRI framework involved the integration of data from two core operational sources: load performance records and shipment summary data. The dataset covered U.S. domestic truckload operations across thousands of active carrier-lane combinations.

The machine learning model—built using Python with SQL-based data extraction—was designed to accomplish four tasks: identify the factors most predictive of pickup defects; quantify the direction and magnitude of each factor's contribution; generate a composite risk score for each carrier; and flag carriers requiring immediate intervention.

For developing the PRI framework, extensive research was conducted to identify the types of data required for performing the holistic analysis. Data related to vendor pickup, weather, load scheduling, carrier historical performance, load characteristics, equipment used, operational site characteristics, and a variety of other input parameters was collected. A total of 147 input parameters were initially used for prediction. A machine learning model was built, and the first step was to improve the model’s accuracy in predicting defects using a real-life dataset. The model was trained on a portion of the data to predict outcomes from the remaining set of data. The model’s accuracy was initially close to 53%. After discussing with domain experts and eliminating non-essential predictors, the model’s accuracy continued to improve. At the end, around 20 input factors were identified to be the most important in predicting the output—in this case, pickup defects. Once these factors were identified and the model was retrained, the model’s accuracy reached almost 85%.

Key predictive factors

The model identified the following as the most significant predictors of pickup defects, ranked by contribution:

Some of the findings from the model that supply chain professionals may find interesting are:

• Pickup time at the supplier site mattered a lot. Night shift pickups showed a higher rate of delays compared to day shift pickups. Especially, pickups after midnight showed a very high rate of defects. While it is not possible to only schedule pickups during day shifts, this pointed toward an opportunity to schedule specific loads during day shifts.
• Specific geographic regions showed higher pickup defect rates compared to others. This suggests that lane-level and origin-specific interventions might be effective in preventing a significant percentage of defects.
• During the study, a number of different driver type deployments were studied, and some of them showed higher defect rates. For critical lanes and customer-sensitive shipments, it is possible to use the low-risk driver type to prevent defects.
• Contract types matter significantly. Contracted carriers with long-term contracts are less likely to be late to pick up a load compared to short-term contracted carriers.

Risk score distribution and model performance

For the purpose of this study, the PRI was developed to generate scores ranging from 0 to 100. Scores below 25 indicate low risk; 25-60 indicate medium risk; above 60 represent high-risk carriers requiring immediate engagement. In the pilot study, the average carrier score was approximately 24.3, indicating that the majority of the carriers performed well. Carriers exceeding a score of 60 were 3.5 times more likely to experience pickup defects compared to the network average. For supply chain professionals who want to adopt this model, the risk tolerance will depend on the specific use case of this model and also on the cost of the defects that will be analyzed.

The scatter plots were developed for a sample of the carriers representing the overall population using Python pandas and seaborn libraries (Figure 1). The top chart shows the relationship between PRI score and pickup defect rate. It shows that high-risk carriers generate a disproportionate number of defects. The bottom two charts show slight inverse relationships between PRI score and cost savings, and PRI score and model detection accuracy.

Figure 1. Carrier Risk Score vs. Pickup Defect Rate, Projected Cost Savings, and Model Detection Accuracy across 320 carriers. Color indicates risk tier. Dashed lines show smoothed non-linear trends. Red dotted line marks the critical threshold (score > 61.9). Note: Specific data points have been anonymized and figures are illustrative of observed patterns.

Two interesting features of the data can be identified here. First, the trend is non-linear and defect rates are relatively low at low PRI scores. Defect rates rise sharply above 40, and cluster at and above 61.9. This means that prioritizing high-risk carriers for targeted actions will result in preventing a high number of defects. Second, the model's accuracy is highest for low-risk carriers and truly high-risk carriers. The middle band is where the model's accuracy is not as high, and may require human intervention to further investigate specific issues related to the carrier pool.

The PRI framework – How does it work?

The PRI model operates as a real-time risk score projection tool. The risk scores for carriers are updated with the latest data almost in real-time as they execute and complete loads. During the pilot study, the model was not automated and the risk scores were manually updated. However, the long-term vision is to have the model automatically ingest new data and learn from that data, so the model’s accuracy improves and the output data is also more actionable in preventing future defects. The PRI model already helps managers to take proactive actions. A more accurate model will enable greater impact from the actions.

Core system components (long-term vision)

• Data integration layer: In this layer, the model will extract data from a variety of load performance and shipment-specific sources and will consolidate them into a singular analytical dataset.
• Machine learning scoring engine: In this layer, the trained ML model weighs the input parameters and generates carrier-specific risk scores.
• Risk stratification dashboard: This layer is a visual interface that segments carriers by risk tiers, highlighting contributing factors and tracking score trends over time.
• Intervention workflow: This feature allows the model interface to send automated alerts to defect owners when a specific threshold is exceeded.

The model enables assessment of risk related to pickup defects and supports proactive intervention to prevent defects. This also allows labor planning teams and operations teams to adjust their plans according to the risk of late shipments.

Demonstrated results

Industry context: Where this fits

While the PRI framework is consistent with the broader industry trend of enabling proactive measures through AI and automation, the application of machine learning in pickup defects is not well researched. Regardless, some similar research efforts are worth mentioning:

• FourKites, project44, and some similar visibility platforms have proved that real-time and predictive ETA capabilities reduce defects, but these tools generally focus on in-transit events. They don’t mitigate pickup risks.
• Academic research on freight market dynamics has highlighted that carrier behavior is heavily dependent on contract type and capacity conditions. The PRI model identifies both as significant factors in predicting pickup defects.
• The 2024 State of Logistics Report (Council of Supply Chain Management Professionals) stated that transportation cost and service reliability are top operational concerns for shippers. The PRI model attempts to solve these two issues through predictive risk management.

The distinguishing feature of the PRI approach is its integration of multiple datasets from different sources across the supply chain and logistics network, which other models fail to achieve—and therefore lack the true predictive nature required for proactive defect prevention. Most existing tools only optimize for one or two factors responsible for on-time performance defects; the PRI approach synthesizes multiple factors to produce a more realistic and accurate risk picture.

Recommendations for supply chain practitioners

An important clarification to note is that the PRI approach is just one use case of how it can be applied to carrier performance management. In theory, the same approach should be applicable to multiple defect types. Based on learnings from the pilot study, a list of recommendations has been compiled for supply chain executives considering a predictive risk approach.

1. Start with defect concentration analysis

Before building a predictive model, analyze how defects are distributed across the carrier base. If a small percentage of carriers account for a disproportionate share of defects, a risk index approach might be what you need. If defects are not concentrated, the problem may be systemic and carrier performance may not be the primary root cause.

2. Prioritize data integration into the model

The PRI model works because it gathers data from a variety of signals across the supply chain. The more accurate input data that the model receives, the more accurate its output will be. Ensuring that the data exists, and if not, measuring data points and making them available to the model is a pre-requisite of the PRI approach.

3. Use risk scores to change the process and mindset

The value of creating risk index scores is the targeted actions that the scores enable. The specific factors that the model identifies make it easier to address concerns and take preventive actions. The team and company will have to trust the model's output once its accuracy is established, and be willing to take proactive actions.

4. Design for real-time update and automation

The PRI model was not fully automated during the pilot study, but supply chain practitioners should target building a fully automated version of the model requiring no manual intervention. Planning, designing and validating a fully automated model early is essential for scaling the system across a full supply chain network.

5. Expand gradually and validate rigorously

The initial model should focus on a small subset of the network. Once the model’s outputs are validated using real data and accuracy is improved, gradually expand it into testing at a larger scope. Expanding too quickly may reduce stakeholder confidence and complicate model refinement.

Conclusion

Managing risk in supply chain has been mostly reactive, but it is slowly transforming toward a more proactive risk management approach. Today, we have the tools, data and processes to identify risks before they become expensive, chronic and unmanageable.

The Predictive Reliability Index framework described in this article demonstrates that proactively managing pickup defects is possible and a low-hanging fruit in some cases to reduce transportation cost.

The model’s accuracy and the ability to reduce defect rates and improve on-time performance provide a compelling case for investment.

More broadly, the PRI approach underscores the mindset that with available data, we should be able to predict every defect in the supply chain network. Fixing a defect proactively not only reduces cost, but also improves team morale. This is the mindset that keeps supply chain practitioners relevant and will keep their business competitive in tomorrow’s increasingly complex supply chain network.

About the author

Debanshu Sharma is a senior supply chain and transportation analytics professional with expertise in predictive modeling, transportation network design, inventory placement, carrier performance management, and large-scale freight network optimization. He regularly publishes in well-known trade publications and serves as a reviewer for peer-reviewed supply chain journals.

Disclaimer: The views expressed in this article are the author's own and do not represent the views of any employer, past or present. All operational data referenced in this article has been anonymized to protect confidentiality. Quantitative figures—including model performance metrics, defect reduction rates, and projected savings—are scenario-based estimates derived from operational analysis and are presented for illustrative purposes only. They should not be interpreted as formally audited or externally validated results.

Executive takeaways

The NextGen Supply Chain Conference is focusing on what matters for supply chains more than ever in 2026: execution. As part of that shift, the conference has revamped its annual awards program, aligning categories more closely with how supply chain transformation is actually happening today and opening submissions for organizations ready to demonstrate real results.

Submissions are now open with a deadline of May 15, 2026.

This year’s awards are designed to move beyond theory, recognizing companies that are deploying technology, scaling operations, and delivering measurable outcomes in real-world environments.

This year’s conference will take place Oct. 21-23, 2026, at the W Nashville hotel in Nashville, Tennessee.

A sharper focus on execution

The updated awards program reflects a broader industry shift. AI, automation, and digital transformation are no longer future concepts, they are operational realities. The question is no longer what’s possible, but what’s working. The 2026 awards are structured across three main areas with more of a focus on execution.

As in the past, we will be honoring both End User companies and Solution Provider firms.  End User Awards recognize the companies that have moved beyond pilots into production environments. The Solution Provider award will honor those companies whose solutions are not just innovative, but actively delivering results for customers across industries.

Both End Users and Solution Providers will be honored in the following categories:

• Intelligent Transformation Award. Recognizing organizations that have embedded AI and advanced technologies into day-to-day operations at scale.
• Autonomous Operations Award. Honoring companies deploying robotics and automation to drive measurable improvements.

Special Recognition Awards

• Visionary Award. The conference’s highest honor, recognizing holistic supply chain leadership.
• Startup Award. Spotlighting emerging companies with strong differentiation and early traction
• Partnership in Execution Award. New for 2026

The submission process for the 2026 awards is now open. Eligibility criteria for each award is available on the NextGen Supply Chain Conference awards page. Submissions are due no later than May 15, 2026. To submit for an award, click here.

A new award for a new reality

The most notable addition this year is the Partnership in Execution Award, which recognizes something the industry has long talked about but rarely celebrated directly: collaboration that delivers results.

This award honors joint achievements between end users and solution providers or consulting partners, focusing on measurable outcomes, shared accountability, and execution excellence.

In an environment where no company operates in isolation, the ability to execute across organizational boundaries is increasingly becoming a competitive differentiator.

The Partnership in Execution Award winner will present during the conference in a fireside chat format. Participation of both the solution provider/consultant and end user company is required for the fireside chat.

More than recognition, it’s participation

Award winners don’t just receive recognition, they take part in the conference program. Recipients are required to attend the NextGen Supply Chain Conference, October 21–23, 2026, at the W Nashville in Nashville, TN, where they will present their work as part of the agenda.

That structure reinforces what NextGen has become known for: a practitioner-driven event focused on real-world execution, not theoretical discussions.

The conference targets approximately 250 senior-level supply chain, logistics, procurement, and operations leaders, with a strong emphasis on peer learning and applied insights.

How to get involved

Organizations interested in participating in the 2026 NextGen Supply Chain Conference have multiple ways to engage:

• Submit for an award (deadline: May 15). Click here.
• Register to attend. Click here.
• Apply to speak by proposing a session focused on real-world challenges and outcomes. Click here.
• Explore sponsorship opportunities to connect with a highly targeted executive audience. Click here.

The 2026 theme—Innovate. Upskill. Transform.—reflects the conference’s continued focus on the intersection of technology, talent, and execution.

Executive takeaways

A term often used in discussions to improve supply chain performance is SKU segmentation. But what does that mean in real-world, practical application?

While every product a company makes and sells can be considered important, in reality each item or SKU serves a different purpose in the company’s product portfolio. Therefore, to optimize total portfolio results, it can be very helpful to divide the portfolio into more similar subgroups. Because these subgroups together equal the whole, they are commonly referred to as segments. The practical use of segments is to clarify the differing role the items or SKUs in each segment play in the portfolio. Then, strategic choices can be made and operational policies enacted by segment to maximize portfolio results. 

Examples of typical segments include trial sizes/market entry products or otherwise strategic items, high volume items, high profit items, ‘fill out the line-up’ items that meet a specific customer or consumer niche, seasonal/promotional items that need responsiveness, and low/slow moving items—these exist in the portfolio but may not have a role in the future. The right number of segments will vary by company and category considering the breadth of the portfolio and business needs. Generally, fewer is better, and most companies can manage with about four or five segments.

Once the role of the segment and items in that segment are defined, strategic policy choices can be made. For example, seasonal/promotional items may need a very responsive supply chain design (shorter lead times, closer suppliers, instantaneous capacity.) Strategic items may carry higher inventory levels to prevent out-of-stocks. High-volume and high-profit items may have more installed capacity, representing their importance to overall portfolio results. The low/slow-moving segment typically will have lower service targets and inventory levels reflecting its overall lower priority. Done correctly, segmentation helps both set up supply chains to match the way the business views SKU priorities, but also to manage basic cost/service/cash trade-offs and to align overall supply chain KPI’s.  

Two things to be aware of: (1) it can be easy to confuse segmentation and the use of product families.  Families aggregate items that are similar (e.g., common production resources or financial profiles).  The purpose of families is management simplification (ability to look at meaningful higher-level aggregate numbers vs. individual items/SKUs). This is not the same as segmentation, which is disaggregation to enable strategic policy differentiation. (2) Organization dynamics often will lead to someone (typically in sales or marketing) being able to make a case for each SKU being strategic. In a typical portfolio (>100 SKUs), it is not realistic for each SKU to be strategic to the same level—this leads to avoiding making prioritization choices on trade-offs needed based on supply chain capabilities. 

Why is segmentation important? The power of segmentation is converging everyone in the organization (commercial + supply chain) to understand the role of the item or SKU in the portfolio. This convergence allows management strategies and decisions to be consistent with the items segment, as well as all supporting planning and operational systems. For example, on POME items, a business should never be out of stock—this is where your strategic target is first entering and shopping your category. If your product is unavailable, they may choose a competitive brand and never give your product a chance.  Therefore, even though trial-size items are ‘starters’—typically young or new to the category with lower price points and not large volume or margin, they are very strategic (examples: teen cosmetics or lower-priced cars for first-time buyers). So for trial items, service is king—and operational policies like inventory and service target setting will be developed to ensure high in-stock levels. Other metrics, like OEE, might be deprioritized for trial-size SKUs. Where segments are coded in planning systems, the segment priority and policies are factored into recommended plans. 

Segmentation enables focus. For example, it’s typical to have differentiated service targets—higher targets on more important segments, lower on less important. This helps the organization deliver the best results where they matter most. Similarly, segmentation helps identify choices on where to invest more inventory—higher safety stock on more important items, lower on less important items. This mindset carries across the supply chain: a business should expect and tolerate lower OEE on trial sizes since service is the priority.

For analogy, an example of segmentation in the service industry is different airline classes of service. First Class passengers are at a higher margin, and as such, get special treatment. All airline systems advise personnel who the first class passengers are and how to provide differentiated treatment (ranging from lounge access, priority security screening, priority boarding, differing on-board experience and service levels, etc.)

Benefits of segmentation: The overarching benefit of segmentation is strategic and operational congruency. The treatment of an item is consistent with its segment, defining its role and importance in the portfolio. The benefits of segmentation show up in areas like:

• increased sales revenue and profit by focusing more on important items, less on the others
• capital and capacity analysis matching business importance
• organization alignment and common focus
• setting service and inventory targets that match the importance to the business vs. one size fits all.
• ability to hard-code segmentation into planning and operational systems to generate recommendations consistent with segment choices.

Watchouts: Unfortunately, there can be many intended or unintended barriers to segmentation, including:

• treating segment as a synonym for product category—it is an item/SKU level assessment
• not having full lead-team participation and treating it as a supply chain only exercise
• static models. Especially at the item level, portfolios change, and the role of items in the portfolio can also change. Segmentation needs to be checked quarterly, renewed at least annually (best practice starting is semi-annually).
• Analysis paralysis. Start light, start quick, learn and adjust. 

How to implement segmentation?

Segmentation is a business team decision, and should be implemented at the multi-functional lead team level sponsored by the general manager (or whoever has P/L responsibility). Particularly if there are current gaps vs. targets (e.g. volume, profit, service, inventory), segmentation should be viewed as an enabling tool to increase focus and improve results where they make the biggest difference.

About the authors

Andrew Byer is a former P&G Supply Chain Leader.  Mike Dobslaw leads EY’s Supply Chain Planning Practice. To learn more about how EY and P&G team to support supply chain transformations please write michael.dobslaw@ey.com

Executive takeaways

Editor’s note: This article first appeared on the University of Tennessee, Knoxville’s Global Supply Chain Institute’s blog. It is being reprinted with permission. You can read the original post here.

Author’s note: In July 2025, the University of Tennessee Global Supply Chain Institute published “Reshoring the Workforce: Bridging America’s Manufacturing Talent Gap,” by Alan Amling and Darrell Edwards. The white paper summarizes research by the Advanced Supply Chain Collaborative, a think tank partnering faculty experts with industry professionals to enhance business performance, enrich teaching, and cultivate supply chain excellence.

In boardrooms across America, the calculus of global manufacturing is changing fast. The pandemic revealed the fragility of long, complex supply chains. Rising geopolitical tensions and the tariff uncertainty have accelerated a strategic rethink. Now, from semiconductors to EV batteries, companies are making multi-billion-dollar bets on American soil.

But an unanswered question looms: do we have the people to power this revival?

The answer, based on recent research conducted by the University of Tennessee, Knoxville’s Advanced Supply Chain Collaborative, is a troubling “not yet.”

The great skills mismatch

The concept of reshoring seems to be an intuitive and smart one: bringing critical manufacturing back home to reduce risk, boost resilience, and create jobs. Government incentives, such as the CHIPS Act and the Inflation Reduction Act, have helped fuel a record $266 billion in U.S. greenfield manufacturing projects in 2024 alone. But this investment surge risks outpacing the current labor market’s ability to respond.

In our survey of supply chain leaders, only one-third of companies surveyed had made any serious move toward reshoring. A commonly cited barrier was a shortage of skilled labor. And for those companies that have reshored, the experience confirms the challenge: the gaps aren’t just in advanced tech or coding, they’re in supply chain operations, leadership, and good old-fashioned problem-solving.

It’s ironic: just as political momentum and market pressures align to bring production home, we’re staring down a workforce wall.

What’s driving the disconnect?

The U.S. hasn’t been a manufacturing powerhouse for decades. As jobs moved offshore, the pipeline of skilled workers, such as machinists, toolmakers, and industrial engineers, dried up. Trade schools withered. Career and technical education lost prestige. Today, we’re paying the price.

Alternatively, automation has transformed the requirements of modern manufacturing. It’s no longer about brute labor; it’s about interdisciplinary thinking. Companies need workers who can manage complex supply networks, apply quality standards, and operate alongside robotics and AI systems. Yet many educational and training systems are lagging behind.

And looming over it all is the demographic cliff. According to the U.S. Chamber of Commerce, as of June 2025, there were 8 million job openings, but only 6.8 million unemployed workers. The fertility rate has dropped to 1.62—well below the replacement threshold—and labor force participation remains stuck below pre-2000 levels. Absent a major policy shift, immigration will be the only source of population growth by 2040.

Reshoring is possible. But without the workforce, it will be expensive and unsustainable.

Strategic risk—and strategic opportunity

Let’s be clear, this isn’t just a talent gap. It’s a strategic risk.

Consider the U.S.-Mexico-Canada automotive supply chain, now facing multiple tariff hits on cross-border parts. Or the $20 billion semiconductor fabs that may take years to become operational, and longer to reach full staffing. These are long-term bets. If we don’t close the workforce gap quickly, the value proposition will evaporate.

But there’s good news. Innovative companies are already showing what works.

Those who have successfully reshored have told us they rely on a combination of internal upskilling, partnerships with technical schools, and targeted executive education. Internships and co-op programs help build pipelines, while apprenticeships bridge the experience gap. Cybersecurity and data analytics training are becoming increasingly important. And almost universally, respondents emphasized that soft skills, such as communication and critical thinking, were equally as important, if not more important than technical skills.

The message is clear: this isn’t about bringing back 1970s factory jobs. It’s about preparing for a new, more sophisticated industrial economy: digitized, automated, and interconnected.

What business leaders can do

For executives considering reshoring, workforce strategy must be integrated into capital strategy, and both must be aligned with corporate strategy.

Based on our research, here are five actions to take now:

1. Assess your skill gaps now. Use capability diagnostics to understand what your workforce can do today and what it will need to do after reshoring.
2. Invest in applied training. Support programs that offer hands-on experience, from co-ops to apprenticeships to industry-led bootcamps.
3. Collaborate with universities to upskill employees. Leverage online learning, executive education, and customized programs to ensure future leaders attain the diverse, critical thinking skills that reflect your real business needs.
4. Elevate leadership and communication. Don’t underestimate the importance of change management. Reshoring is a cultural as well as operational shift.
5. Make workforce strategy a boardroom issue. Labor availability isn’t an HR problem—it’s a business continuity risk.

This isn’t just a large-company game. Small and mid-sized manufacturers, with the right partnerships and incentives, can move more quickly and nimbly to build tailored talent pipelines. In many cases, their size can be a strategic advantage, allowing for rapid alignment between leadership and workforce needs.

Where policy must step in

Businesses can’t do it alone. Immigration reform, especially around high-demand industrial skills, must become a national priority. Expanding pathways for vocational education—through tax incentives, Pell Grant eligibility, and performance-based funding—can help close the gap more quickly. And federal programs should reward companies that invest in their people, not just in their factories.

We’re in the early innings of a U.S. manufacturing revival. But unless the workforce dilemma is resolved, that revival will stall before it gains momentum.

Final thought

The U.S. is rediscovering something many had forgotten: Industrial strength isn’t just about machines, it’s about people. If we want the future of manufacturing to be made in America, we need to make sure America is ready to make it.

To learn more about how your company can partner with the University of Tennessee Global Supply Chain Institute to explore advanced concepts in supply chain management, visit ASCC.

Download the white paper to read more about how supply chain leaders can bridge the talent gap in reshoring manufacturing.

About the author

Alan Amling is a TED speaker and thought leader on harnessing digital disruption for success. Currently a lecturer and fellow at UT, Amling helped drive innovation over a 27-year career with UPS. He is the CEO of the advisory firm Thrive and Advance LLC and serves on the executive advisory board for the Georgia Tech Manufacturing Institute. He researches, invests, advises and speaks on innovation in manufacturing and logistics and on how firms can recognize and thrive in disruption.

Amling’s last role at UPS was as vice president of corporate strategy, where he helped revitalize the company’s innovation and venture capital programs for the digital economy. He moved into this role after serving as VP of marketing for UPS Global Logistics & Distribution.

Executive takeaways

A secure industrial supply chain is no longer defined by physical controls around a defined network perimeter and contractual safeguards with suppliers. In an era of maturing artificial intelligence, open-source software, interconnected vendor ecosystems, and increasingly sophisticated cyberattacks, supply chain resilience also depends on two foundational elements: behavioral (including identity) data and layered artificial intelligence.

Cyberattacks are on the rise across every industry, but manufacturers and their partners are especially vulnerable due to the nature of supply chain interconnectedness. One compromised system can lead to the infiltration of all dependent systems, software applications, cloud instances, and networks. Today’s attacks employ diverse methods that escape many common security measures and detection tools.

What keeps many CISOs up at night is the pressure to stay on top of the increasing information security and privacy threats that industrial organizations face on a daily basis. This is where contextual threat data and a layered AI defense can thwart sophisticated supply chain attacks across the unified IT/OT network.

The expanding industrial attack surface

Manufacturers, energy providers, and critical infrastructure operators increasingly rely on digital systems to manage procurement, logistics, production, and distribution. Enterprise IT networks now routinely intersect with operational technology (OT) environments, including industrial control systems and SCADA platforms. This unification has many benefits, including visibility and efficiency, but the downside to this connectivity is vulnerability.

High-profile incidents such as the Colonial Pipeline ransomware attack and the SolarWinds software supply chain compromise demonstrated how security weaknesses can cascade across industries. A single compromised vendor, credential set, or software update can significantly disrupt operations at scale.

Making matters more urgent, threat actors can now use AI to help them scan for exposed assets, generate highly convincing phishing campaigns, and probe vendor ecosystems at machine speed. Traditional, perimeter-based defenses are no match for this level of automation.

The good news is that defenders can also leverage AI to fight back against these threat actors.

Network detection and response: Using data and AI to flag anomalies

Network detection and response (NDR) is an emerging category of cyberdefense that’s transforming how industrial organizations build more resilient supply chains. NDR is especially relevant for manufacturing and other industrial organizations because it has the power to detect suspicious identity and user behaviors early, before a cyberattack can cause significant harm.

Early detection of cyber intruders is critical to the manufacturing supply chain because threats in this sector don’t stay contained in a server. They ripple into physical production, logistics, revenue, and even safety. In manufacturing, minutes matter. The longer a threat actor stays undetected in your systems, the more potential for chaos.

Unlike endpoint detection and response (EDR), a defense strategy that focuses on protecting individual devices, or endpoints, NDR monitors and analyzes all network traffic across the entire environment—including IT, OT, and cloud networks—to detect malicious activity. It goes beyond the devices themselves to look at the behaviors.

Manufacturing networks are complex. ERP systems, supplier portals, plant-floor controllers, IoT devices, and remote vendor connections are interconnected, each requiring login credentials. The problem is that every set of credentials introduces another point of failure where a threat actor can gain access. NDR monitors east-west and north-south traffic, giving security teams insight into lateral movement between corporate and plant networks. Communications between industrial control systems (ICS) and vendor remote-access sessions can be patrolled for any unusual traffic to cloud-based logistics platforms.

Because NDR analyzes network metadata and behavior rather than relying solely on endpoint agents, it is particularly valuable in OT environments where agents cannot easily be installed.

NDR uses AI to quickly identify anomalous patterns and indicators of compromise (IOCs) in the supply chain that would otherwise go unnoticed by traditional security measures. That’s because a majority of today’s supply chain attacks stem from compromised user credentials and account takeovers—invisible intruders masquerading as legitimate users. However, these stealthy cybercriminals usually have “tells” that advanced behavior analysis can detect. 

Examples of these “tells” include unusual network traffic patterns to or from OT devices, such as PLCs and SCADA systems; unexpected or unauthorized attempts to access external IP connections from OT systems or unauthorized protocols on OT networks, such as SSH or RDP on a controller. Sometimes, it’s as simple as an unusual or unauthorized change in control logic or firmware on a connected device, or multiple, failed logins from unexpected locations, times, or user accounts. Other red flags include the attempted use of default, generic, or expired credentials, new user accounts that suddenly appear on OT systems, and equipment that suddenly behaves erratically or inconsistently without a mechanical cause.

AI-driven NDR solutions have the ability to detect these anomalies and more. By analyzing live network traffic across the manufacturing supply chain, NDR provides deeper visibility into potential cyber threats, uncovering malicious activity that often slips through the cracks of traditional security measures. These insights enable security teams to rapidly contain and neutralize threats before they can cause widespread damage.

Why layered AI matters

Using AI to defend the supply chain from cyberattacks should not be a single, monolithic system. To work properly as a defense strategy, AI must be layered across detection, correlation, and response. There are multiple types of AI that can work together to assist security analysts, helping them find threats and act faster to stop them.

1. Detection layer: Machine learning

Machine learning models automate the most basic monitoring functions and the detection layer of the network. Here, AI can identify anomalies across user behavior, device activity, network traffic, and supplier interactions. In industrial environments, this includes deviations in production workflows or command sequences within OT systems. The detection layer is the first line of defense against zero-day threats or unknown attack vectors.

2. Correlation layer: Graph ML

Think of this layer as where the dots start to connect. AI takes the anomalies discovered in the detection layer and correlates signals across domains—linking a suspicious vendor login to unusual lateral movement inside a plant network, for example. This cross-domain synthesis reduces alert fatigue and prioritizes material risk.

3. Response layer: LLM and agentic AI

This layer is where innovation is really happening. Automation, in the form of large language models (LLMs) and AI agents, can assist human analysts by triggering a specific response to threats. This can happen in a few different ways. With an LLM-based response, an AI assistant provides the human analyst with instructions on the next step, such as isolating compromised systems, revoking credentials, or triggering supplier risk workflows. With an AI agent, very little human intervention is needed, as the agent carries out the next set of actions.

Layered AI shifts security operations from reactive investigation to proactive risk management. It also helps industrial firms cope with a persistent cybersecurity talent gap by automating processes, eliminating false positive alerts, and pulling together context for human security analysts.

Human oversight remains critical

Despite advances in AI, supply chain security cannot be fully autonomous. Industrial systems have safety implications, regulatory constraints, and operational nuances that require experienced judgment.

AI should augment—not replace—security operations teams. Analysts must validate AI-driven conclusions, investigate root causes, and coordinate with suppliers when incidents occur. The goal is not a “lights-out SOC,” but a human-augmented SOC model that scales expertise.

Industrial organizations that invest in this cyber defense strategy will not eliminate all security risks. But they will dramatically reduce dwell time, make their supply chain systems more resilient, and maintain operational continuity.

About the author

Subo Guha is the senior vice president of product management at Stellar Cyber, where he spearheads the development of the company's award-winning, AI-driven Open XDR solutions. With over 25 years of experience, Guha has held senior leadership roles at industry-leading companies like SolarWinds, Dell, N-able, and CA Technologies.

Executive takeaways

After more than 15 years architecting forecasting algorithms, leading S&OP transformations, and navigating the operational realities of supply chain networks, one guiding principle has remained constant: a forecast is only as valuable as the decisions it improves. This truth has become even more important as organizations adopt increasingly complex AI‑driven forecasting architectures.

The last few years have brought an explosion of interest in AI agents, large language models, and open‑source machine learning frameworks. These tools offer unprecedented analytical power, yet they also introduce new layers of complexity. Without a disciplined mechanism to evaluate whether each new model, feature, or agent actually improves outcomes, organizations risk building forecasting environments that are technologically impressive but operationally ineffective. This is where forecast value add (FVA) becomes indispensable.

Why FVA matters in today’s AI landscape

Forecasting conversations have shifted dramatically. Where organizations once debated ARIMA parameters or exponential smoothing coefficients, they now grapple with questions about transformer‑based time‑series models, multi‑agent orchestration, external signal ingestion, and LLM‑driven feature engineering. The sophistication of the tools has increased, but the fundamental challenge remains unchanged: Leaders must determine which innovations genuinely improve accuracy and which simply add complexity.

FVA provides the clarity needed to answer that question. It quantifies the incremental contribution of each forecasting step, allowing organizations to distinguish meaningful innovation from noise. In an era where forecasting pipelines can easily become over‑engineered, FVA ensures that every component earns its place.

The realities of e-commerce outbound‑flow forecasting

Outbound‑flow forecasting is where forecasting errors become painfully visible. In e-commerce environments, even small inaccuracies can cascade into operational disruptions. When forecasts miss the mark, labor schedules become misaligned, carrier capacity is incorrectly allocated, inventory is misplaced, split shipments increase, and service levels decline. These failures ultimately erode customer satisfaction and increase cost‑to‑serve.

Forecast Value Add Case Study at NextGen

 

Agentic AI in Action: Reinventing Demand Forecasting at Scale

A real-world case study on how a high-velocity retailer re-architected its eCommerce forecasting using agentic AI orchestration and a revitalized forecast value add (FVA) discipline to drive measurable accuracy, faster response to volatility, and scalable operational impact.

Join Bijoy Sasidharan at the 2026 NextGen Supply Chain Conference, Oct. 21-23, 2026, at the W Nashville hotel in Nashville, Tennessee, where he will present case studies on how FVA has improved the forecasting capabilities of Fanatics. 
 

Register today

 

Outbound forecasting extends far beyond predicting demand. It requires anticipating where, when, and how demand will convert into physical movement across a distributed network. The complexity is amplified by multi‑node fulfillment architectures, same‑day and next‑day delivery commitments, promotional spikes, high return volumes, channel blending, and rapid SKU proliferation. These realities have accelerated interest in AI agents and orchestration frameworks capable of handling scale, heterogeneity, and real‑time decisioning.

How agentic AI framework such as LangChain are reshaping forecasting

Organizations are increasingly deploying forecasting algorithms that behave like specialized digital analysts. These agents autonomously pull data from multiple systems, execute diverse forecasting models, engineer features, detect anomalies, explain deviations, generate scenarios, and collaborate with human planners. LangChain serves as the orchestration layer that coordinates these capabilities. It manages tool routing, memory, context, multi‑step reasoning, and communication between agents, effectively functioning as an operating system for forecasting workflows.

However, more agents do not automatically translate into better forecasts. Without FVA, organizations may unintentionally introduce redundant or counterproductive steps. FVA becomes the mechanism that ensures each agent contributes measurable value.

Building an FVA‑driven AI forecasting pipeline

A mature, AI‑enabled forecasting pipeline typically includes several structured components. Three of these—FVA computation, human‑in‑the‑loop review, and continuous learning—are especially critical.

1. FVA computation

FVA computation becomes the analytical backbone of the forecasting pipeline. For every SKU‑location‑day combination, the system evaluates error metrics such as MAPE, WAPE, RMSE, and bias to understand both magnitude and direction of error. It then compares the performance of each forecasting step against the baseline model, against human overrides, and against previous model versions. This layered comparison reveals whether a particular agent, adjustment, or model iteration genuinely improves accuracy or inadvertently introduces noise.

Over time, these evaluations create a transparent performance record that guides investment decisions, model selection, and process refinement. Instead of relying on intuition or anecdotal evidence, leaders can see precisely which components consistently add value and which require redesign or retirement. FVA transforms forecasting from a model‑centric exercise into a value‑centric discipline.

2. Human‑in‑the‑loop review

Human planners remain essential, but their role evolves significantly in an AI‑driven environment. Rather than manually adjusting forecasts based solely on experience or intuition, planners receive a structured package of insights that includes forecast outputs, FVA scores, and natural‑language explanations generated by LLMs. These explanations translate complex model behavior into operational language, enabling planners to understand why a forecast changed and what factors influenced it.

This approach elevates human judgment. Planners focus on exceptions, contextual insights, and strategic decisions rather than routine adjustments. When overrides occur, FVA provides accountability by showing whether the intervention improved or degraded accuracy. Over time, this feedback loop strengthens forecasting discipline and reduces unnecessary manual adjustments.

3. Continuous learning

An AI‑driven forecasting system must evolve with the business. Continuous learning mechanisms ensure that the system adapts to shifting demand patterns, promotional cycles, and macroeconomic changes. Drift detection identifies when historical relationships no longer hold. Automated retraining keeps models calibrated without requiring manual intervention. Reinforcement learning allows agents to improve based on historical performance, while self‑evaluation mechanisms enable each agent to assess its own contribution relative to peers.

Together, these capabilities create a self‑improving ecosystem. Instead of degrading over time, the forecasting system becomes more accurate, more resilient, and more aligned with operational realities.

Industry perspectives from the field

Across my career, I’ve had the opportunity to work closely with organizations in omnichannel retail, life sciences and apparel e-commerce. While these sectors differ dramatically in their products and operating models, they share a common truth: forecasting is never just a mathematical exercise. It is a reflection of how demand behaves within each ecosystem, how decisions flow through the network, and how operational constraints shape what is possible. Each industry presents its own forecasting challenges, and understanding these nuances is essential for designing AI‑driven systems that actually work.

Omnichannel retail: Complexity at every node

In omnichannel retail, the most striking challenge is the interconnectedness of the network. Retailers operate stores, distribution centers, fulfillment centers, dark stores, and same‑day delivery hubs—each with its own demand signals, constraints, and service promises. The difficulty is not simply forecasting demand but forecasting how that demand will route itself through a constantly shifting network. Online traffic, in‑store footfall, local events, weather patterns, and promotional calendars all influence where orders originate and how they should be fulfilled. Behaviors such as store pickups or ship‑from‑store blur the line between digital and physical demand, making it difficult to treat channels as independent entities.

What I’ve consistently seen is that when organizations attempt to forecast each channel in isolation, misalignment becomes inevitable. Inventory pools drift out of balance, stores become overloaded with fulfillment tasks they weren’t staffed for, and distribution centers struggle to keep pace with last‑mile commitments. The real challenge is not the volume of data but the fragmentation of it. Omnichannel forecasting requires a system that can interpret signals from multiple nodes simultaneously and understand how they interact. Without that, even the most sophisticated models struggle to keep pace with the operational realities of modern retail.

Life sciences: Forecasting with consequences

Life sciences forecasting is fundamentally different because the stakes are higher. In this domain, forecasting errors don’t just affect cost—they can affect patient outcomes. Outbound flows depend on epidemiological trends, provider ordering behavior, cold‑chain constraints, and regulatory requirements. Demand can shift rapidly in response to disease outbreaks, policy changes, or shifts in provider protocols. In my experience, the most challenging aspect is the variability in provider behavior. Clinics and hospitals often order in bursts, influenced by patient flow, reimbursement cycles, or inventory management practices that differ widely across regions. These patterns rarely align with traditional time‑series assumptions.

Cold‑chain requirements add operational rigidity; products must move through the network within strict temperature windows, which limits the flexibility of the fulfillment strategy. Regulatory oversight further constrains how inventory can be stored, transported, and allocated. During periods of heightened public health activity—such as flu season or COVID outbreaks—the forecasting system must respond quickly and accurately. Human overrides can be valuable in these moments, especially when planners have access to real‑time clinical insights.

Sports merchandising e-commerce: Volatility as a constant

Forecasting in sports merchandising operates in a level of volatility few other categories ever face. Demand can shift overnight based on unpredictable moments—an unexpected win, a breakout performance, player trade, a record‑setting play, or a viral highlight. Product lifecycles are short, often measured in weeks, and the window to capture demand is narrow. High return rates further distort the relationship between gross demand and true consumption, making historical patterns unreliable on their own.

From my experience, the hardest part is separating meaningful signals from the noise that surrounds major sports moments. Social media activity can be a strong predictor of demand, but only when interpreted with context. A spike in attention may reflect genuine fan interest, or it may be tied to a moment that never converts into merchandise sales. Treating every spike as a trend leads to over‑forecasting; ignoring them leads to missed opportunities.

Returns and cancellations add another layer of complexity because they create a moving target for net demand. Without explicit return‑adjusted modeling, forecasts consistently overestimate true consumption. Lifecycle dynamics complicate things further. Many items follow recognizable aging curves, but the pace of decline varies dramatically depending on the athlete, the team’s performance, and the timing within the season. The challenge is not just modeling these curves but doing so quickly enough to influence allocation, replenishment, and markdown decisions before the moment passes.

In this environment, timing is everything. A forecast that is accurate but late is often as damaging as one that is wrong.

The patterns that always hold true

Across industries, several lessons consistently emerge. Organizations that start with a simple baseline build a stronger foundation for meaningful FVA measurement. Those measuring every step in the forecasting process gain clarity around where value is created and where it is lost. Automating FVA ensures that evaluation becomes a continuous discipline rather than a periodic exercise. Explainability tools help bridge the gap between data science and operations, enabling planners to understand and trust AI‑generated insights. And despite the power of automation, human judgment remains essential for contextualizing forecasts and navigating ambiguity.

FVA is the compass that keeps AI‑driven forecasting grounded in operational reality. As organizations adopt increasingly complex agent‑based systems, FVA ensures that every step—every model, every agent, every override—delivers measurable value. The future of forecasting belongs to organizations that combine AI automation, human expertise, and rigorous FVA discipline.

About Author

Bijoy Sasidharan is the director of analytics for capacity planning and forecasting at Fanatics, where he leads the development of forecasting models, network capacity planning, and AI‑driven decision tools in one of retail’s most fast‑moving and unpredictable environments. He has held prior roles at Walmart.com, ZS Associates and Kids2 Inc, gaining experience across analytics, supply chain operations, and applied artificial intelligence. With more than 15 years in the field, Sasidharan focuses on building forecasting approaches that are both rigorous and practical, helping teams make better decisions at scale. He holds a master’s degree in supply chain engineering from the Georgia Institute of Technology.

Executive takeaways

Artificial intelligence is not a new technology, but its rapid adoption in the past two years has garnered all the headlines. From chatbots to generative and now agentic AI, the technology is all anyone seems to want to deploy.

But in warehouses, there is another trend quietly taking place, and while AI is part of the story, it is not the full story.

“I don’t know what flipped last year,” said Sankalp Arora, CEO of Gather AI, “but customers have now started to get why they need digitization; why they need solutions like us. If you know what’s happening in your warehouse, that leads to better on-time in-full, better labor productivity, fewer damages, fewer overages, fewer shortages.”

The realization that connecting floor-level visibility directly to OTIF, labor efficiency, and shrink reduction appears to be accelerating adoption of what Arora calls “physical AI.”

Closing the physical-digital gap

For years, supply chains have invested heavily in digital planning systems, control towers, and predictive analytics. Yet a persistent gap remained between what systems believed was happening in a warehouse and what was actually happening.

“There has been a gap between physical and digital thus far,” Arora told Supply Chain Management Review in a recent interview. “What we are doing is we’ll get you the on-the-ground truth of what’s actually happening in the facility.”

Gather AI’s approach uses cameras affixed to assets. Initially that was drones, but advancements have allowed the cameras to attach to other assets such as forklifts. Rather than requiring workers to scan barcodes manually, the system digitizes inventory state passively as the drone, the forklift, and potentially even people, navigate the warehouses.

“The natural extension … is getting moving cameras to collect data,” Arora said. “Things that move in a warehouse are pallet jacks, forklifts … people and other equipment.”

The long-term goal is simple: provide end-to-end visibility from inbound dock to outbound shipment without the manual labor to do physical inventory checks.

Beyond inventory counts

While automated inventory accuracy is the entry point for many customers, Arora says the real value lies in orchestration.

“Once you start getting all of that data, then we know what’s actually happening in the facility better than a lot of tools that they have,” he said. “Then we are very well positioned to orchestrate the facility.”

In practice, Arora said customers typically see ROI through improved on-time, in-full performance; labor efficiency gains and a reduction in overages, shortages, damages and write-offs.

In one case study Arora cited, an inventory control and quality assurance (ICQA) department shrank from six employees to one, while inventory errors dropped 70%.

But Arora pushes back on the idea that this is about labor replacement.

“I don’t think anyone in the warehouse really is replacing labor today because everyone’s short on labor,” he said. “We let people focus on tasks they’re good at and we take away the ICQA part of it.”

AI before AI was cool

Part of the confusion surrounding warehouse AI stems from terminology. Arora is careful to distinguish between different types of AI technologies.

“Before large language models came about, there was deep learning, which was really good for computer vision,” he explained. “Now, once you have that data digitized, then you get into the Gen AI and large language model stuff.”

Computer vision models interpret images and Classical Bayesian AI (an AI system that utilizes Bayesian statistics and probability) governs movement and guarantees in dynamic systems like drones. Generative AI excels at extracting insights and enabling natural language interaction but is “not really good at reading images … [and] not really good at controlling robots,” Arora noted.

In other words, AI is not one thing as many think of it, but rather a stack of technology executing various tasks. Arora notes that the industry has rebranded similar technologies repeatedly: machine learning, big data, AI, generative AI.

“AI existed way before all the marketing dollars were thrown onto them,” he said.

In a market saturated with “AI-powered” claims—some substantive while others are cosmetic—it is the ability to understand the different tools available and choose the right ones for the job at hand.

Hardware agnostic, capital efficient

Another defining element of Gather AI’s model is its hardware-agnostic approach. The company, which in February announced a $40 million Series B funding round led by Smith Point Capital, made a conscious decision early on to build the vision system and not the drones themselves. The software is designed to work with off-the-shelf drones or other assets. It is a decision that Arora believes was in the best interest of the company and paves the way to the best opportunity of success.

“We realized that the tech stack that we have is uniquely positioned to make off-the-shelf hardware work,” Arora said. This allows customers to avoid heavy capital expenditure and Gather AI is able to avoid large capital investments to build factories.

It also allows deployment in extreme environments from -20°F cold storage to some of the hottest warehouses globally. “We work in cold storage at minus 20 Fahrenheit, and we work in the hottest warehouses in the world,” he said.

Cold storage accuracy carries outsized importance, Arora noted.

“If you make a mistake … it’s four times more expensive than an ambient mistake,” Arora said. “It’s more important to get that right without human intervention, because it’s not a human-friendly environment.”

Arora says the latest funding round will primarily support scaling customer operations and expanding modality coverage.

The company currently operates in the U.S. and Dubai, with plans to expand globally.

Executive takeaways

CSCOs are confronting financial mandates to reduce supply chain costs as conflict in the Middle East introduces new constraints and price shocks. This is prompting leaders to review potential increases in transportation costs and reassess previously established cost management strategies.

When it comes to cost management, many CSCOs rely on periodic cost reviews and, often, must deal with urgent enterprise mandates to cut costs. These approaches, however, struggle to identify the complex drivers of cost that exist in supply chains.

CSCOs are often caught unprepared to deal with the volatility of disruptive events that increasingly influence supply chain costs. At the same time, expectations of CSCOs have been rising. Gartner research shows that 73% of CxOs believe CSCOs should drive financial goals.

Gartner predicts that by 2030, 40% of supply chains will use AI to shift toward proactive, data-driven cost management. This reflects a meaningful change in how cost is understood and managed across the enterprise.

AI will help CSCOs avoid broad-brush cost-cutting with more precise cost management. There are three key ways AI can help CSCOs take a more proactive approach to cost management.

1. Revealing cost opportunities hidden in supply chain data

Most organizations already capture extensive data across planning, sourcing, manufacturing, and logistics functions. However, siloed functions often prevent leaders from realizing ways to drive true cost performance.

AI can process large volumes of supply chain data and identify connections that are more difficult to detect through traditional analysis. These insights allow CSCOs to better understand cost to serve across product lines, customers, and channels.

For example, increases in fuel surcharges or marine cargo insurance premiums associated with regional instability may not immediately register as network‑level cost drivers when evaluated independently within procurement or transportation functions. However, when combined with changes in routing, transit times or inventory positioning, these inputs can have a significant impact.

AI‑enabled analysis allows supply chain leaders to detect these cost interactions earlier, helping isolate margin leakage before it manifests on the bottom line. By connecting cost drivers across the full network, CSCOs can gain visibility into savings opportunities that may have previously remained out of reach.

2. Identifying targeted optimization opportunities in real time

Traditional cost-cutting initiatives are often operated as urgent responses to enterprise‑wide mandates. This can result in actions that reduce expense in one area while creating unintended consequences elsewhere. AI enables ongoing scanning of supply chain data to identify targeted optimization opportunities.

For example, conversational AI tools can allow planners to interact directly with supply chain data and ask targeted questions about inventory policies. This allows planners to identify specific SKUs where safety stock policies could be adjusted in response to increased transportation and shipping costs.

Rather than implementing broad inventory reductions, planners can pursue targeted changes that reduce excess inventory without compromising service levels, helping offset higher freight expenses associated with alternative routings or mode shifts.

3. Modeling the financial impact of evolving supply chain conditions

As supply chain structures adjust to changing geopolitical conditions, organizations may face additional costs associated with new material sources, alternative routing, fuel surcharges, emerging markets or the implementation of war‑risk premiums across affected trade lanes. AI‑driven scenario modeling allows CSCOs to simulate how these evolving cost drivers may affect supply chain financial performance under different future scenarios.

CSCOs need to engineer profit by managing costs across the business, not just inside their own walls. This requires incorporating cost visibility directly into planning and decision‑making processes. Digital representations of the end-to-end supply chain can be used to stress test cost initiatives and assess the financial implications of various responses.

Leaders can quantify trade‑offs between cost, service reliability, speed, and agility before implementing changes. This allows supply chain teams to prepare responses to financial mandates in advance and present informed recommendations that align cost management with profitable growth.

From reactive cost cutting to proactive cost leadership

In the current environment, cost management must become an ongoing capability rather than a periodic response to financial pressure. AI‑enabled analysis allows CSCOs to quantify operational trade‑offs and provide strategic guidance that supports enterprise decision‑making as external cost drivers continue to evolve.

Thanks to AI, the CSCO can become a trusted partner to the C-suite by demonstrating trade-offs. With the support of AI, CSCOs can respond to cost-cutting mandates with a strategic “yes, but,” acknowledging the request while providing context on the consequences and offering advice on how to achieve profitable growth.

About the author

Pierfrancesco Manenti is a VP Analyst in Gartner's Supply Chain Practice. He provides insights and advisory support to CSCOs into the future trends and key challenges affecting end-to-end supply chain strategy. Pierfrancesco and other Gartner analysts will provide additional insights on AI and cost management strategies at the Gartner Supply Chain Symposium/Xpo, taking place May 4-6 in Orlando, FL. Follow news and updates from the conferences on X using #GartnerSC.

Executive takeaways

Supply chain technology continues to evolve, and one of the changes taking place revolves around autonomy and the increasing use of artificial intelligence.

“The difference between deterministic and non-deterministic systems is where we’re headed,” said Pano Anthos, founder and managing partner at XRC Ventures, during a recent conversation with Supply Chain Management Review.

For decades, enterprise supply chains were designed around deterministic logic, standard operating procedures, Six Sigma rules, and rigid workflows. “Enterprises are deterministic,” Anthos said. “You will do this and do it this way.”

But agentic AI is challenging that structure, Anthos said. “What would happen if you put AI out there and said, ignore the SOP … don’t break any rules, but get the job done?” he asked.

From rigid rules to adaptive systems

Anthos argues that many corporate rules exist not because they are optimal, but because they were designed for human constraints. Historically, processing power and analysis costs forced simplification, he said.

For example, deciding whether to ship via air or ocean often defaulted to air for simplicity. “You always ship air freight … keep it simple,” he said, noting that an analysis might have turned up a different answer, but the costs to run that analysis didn’t justify doing so.

With AI, computing tradeoffs becomes instantaneous, though, and decisions once considered too complex or time-consuming can now be modeled continuously.

Computer vision and the “blind warehouse”

Beyond decision logic, Anthos sees another major transformation unfolding inside physical operations. “The cost of computer vision is dropping,” he said. “What you’re going to see is more real-time diagnostics capability inside the warehouse and manufacturing floors.”

Warehouses historically relied on visual supervision and periodic audits. “The factory manager sits in a little office off to the side with a little window,” he noted. “Those days are going away.”

Computer vision acts as a diagnostic layer identifying problems in real time. That has implications for safety, insurance rates, and productivity.

The rise of vertical AI

In the investment community, Anthos says the biggest opportunity isn’t foundational large language models, but vertical AI applications.

“There’s a lot of interest in vertical AI right now,” he said. “You can unlock a lot of value and sell a company with very little capital” if it is building vertical AI. He cited an example of a five-person company managing aircraft parts supply with AI for Iberia Airways.

The system, which can be installed in just a day, monitors procurement communications by analyzing email patterns rather than requiring deep ERP integrations. “They sniff email … they attach to your email system and watch traffic between buyers and suppliers,” he said.  

The impact? Anthos said the solution reduced delays by 25%.

Such capital-efficient, domain-specific solutions represent a departure from the traditional “big consulting, big deployment” model. Anthos believes integration barriers have fallen dramatically thanks to API architectures and AI mediation layers.

AI is eating SaaS

Anthos also argued that legacy SaaS platforms are vulnerable because “AI is eating software.”

In essence, he said many providers are adding AI on top of current solutions rather than building natively around AI. The result is that companies are not getting “truly agentic AI [solutions],” he said.

In contrast, AI-native systems operate flexibly, capable of moving across datasets and functions autonomously. He predicts increased pressure on traditional ERP processes such as procurement, freight audit, and CRM workflows where AI can automate root-cause analysis and decision support.

“Freight audit—AI is destroying the freight audit business,” he said.

Instead of merely flagging discrepancies, AI identifies patterns and suggests preventive actions. That shift from reporting “what is” to diagnosing “why it is” is critical, he said.

“I don’t want to know the invoice is 90 days late,” he said. “I want to know why.”

Agentic friction: Cultural and political barriers

Yet while technology advances quickly, corporate culture lags. “Boards are pounding CEOs; what are you doing about AI?” Anthos said. But, executives often respond by installing AI tools reactively rather than strategically. Many projects fail because they don’t address real pain points.

More importantly, agentic systems disrupt political structures. “Mid-level managers don’t like that,” Anthos said of AI surfacing root causes across silos. Autonomous analysis reduces information gatekeeping and challenges hierarchies built around controlling access to data.

A flood of applications

The scale of change may be overwhelming, and quickly.

“There’s a projection that there will be 1.4 million native AI applications by 2030,” Anthos said.

That raises questions about consolidation, sustainability and enterprise readiness. How many freight audit AI systems does the market need, for instance? Or how many procurement agents are really needed?

Investors like Anthos are increasingly looking for capital efficiency and defensible domain expertise. “We’re looking for a team that will break a brick wall down,” Anthos said.

Ultimately, Anthos believes the future of supply chain AI lies in reasoning rather than reporting. “It’s the why and the root cause analysis where there’s real opportunity,” he said.

Artificial intelligence is reshaping supply chain conversations, promising faster decisions, predictive insights, and automated optimization. But AI is only as effective as the data that fuels it. Without clean, structured, and governed data, even the most advanced AI models will amplify inaccuracies rather than eliminate them.

In this webinar, Elenna Dugundji, Director, Deep Knowledge Lab for Supply Chain and Logistics Research Scientist with MIT explores why data discipline remains the foundational driver of supply chain optimization. Our discussion will examine the critical role of data quality, governance, system integration, and process alignment in enabling meaningful AI outcomes. 

Attendees will gain practical insight into how leading organizations are strengthening their data foundations to unlock sustainable optimization—before, during, and beyond AI adoption.

Speaker: Elenna Dugundji, Director, Deep Knowledge Lab for Supply Chain and Logistics Research Scientist, MIT

Moderator: Brian Straight, Editor in Chief, Supply Chain Management Review

Executive takeaways

Procurement has always had a clear mandate: aggregate spend, consolidate suppliers, and drive cost savings. And for decades, that strategy worked fine.

Category managers were rewarded for bundling volume, negotiating rebates, and concentrating spend. But that model was built for a relatively stable geopolitical environment, and since COVID, that environment no longer exists.

In a recent discussion with Supply Chain Management Review, Mark Schenecker, VP of direct materials at Coupa, described how the aggregation-first playbook is colliding with tariff volatility, geopolitical fragmentation, and structural uncertainty.

Aggregation made sense — until it didn’t

Procurement’s aggregation era rewarded concentration. If a company could commit 80% to 100% of a category to a single supplier, it unlocked pricing tiers, rebates, and simplified management. As Schenecker put it, “the mantra that came down was aggregate the spend.”

When disruption was sporadic—a hurricane, a port strike, or a supplier bankruptcy—the risk could be managed tactically. COVID exposed the limits of that thinking, though, and tariffs are accelerating it and showing how much vulnerability is embedded in sourcing models.

Tariffs are no longer episodic

Tariffs have shifted from temporary trade tools used by administration to a global trade policy which has introduced many more variables as rates go up and down, sometimes on a weekly basis. Companies can no longer assume tariff baselines will hold, and even when specific measures are rolled back, pricing rarely resets quickly, if at all.

Schenecker noted that suppliers have responded by embedding tariffs into pricing, forcing procurement teams to change how they model cost. “Landed cost now has to have a discrete item,” he said.

That dynamic forces procurement to rethink the landed cost modeling. Instead of treating tariffs as a background assumption, they must now be discrete, visible line items, Schenecker noted. Companies are increasingly maintaining tariff tables alongside finance, tracking country-of-origin exposure in detail, and modeling timing impacts such as whether duties apply at border crossing, receipt, or bonded release.

If 100% of volume is tied to a single geography, and that geography becomes tariff-exposed, the cost and continuity impact is immediate and severe, Schenecker said.

Geographic diversification is the new leverage

The response is not simply adding a second supplier, he pointed out, but adding a second geography.

The traditional dual-source model often kept suppliers within the same region or even the same country. That model is no longer sufficient in a tariff environment. True resilience now requires geographic diversification to mitigate trade and political exposure, but there are trade-offs that must be considered.

“You can’t give 100% of your spend for a product category to one supplier,” Schenecker said. “You’ve got to keep two, three suppliers and you can’t have ’em from the same geography.”

Volume concentration might drive rebate leverage, but resiliency is put at risk in 2026. Procurement leaders are now being forced to balance cost efficiency against structural resilience.

Network design can no longer sit in isolation

The most significant operational implication of this shift may be organizational. Historically, network design lived within supply chain planning. Procurement executed sourcing strategy based on that design and while the two functions coordinated, they operated on different cadences. Schenecker said that separation is no longer optimal.

“We built for cost. We didn’t build it for resiliency,” he said of the old model.

Now, sourcing optimization, supplier selection, contract strategy, and network configuration are becoming interdependent variables in a continuous model. This integration enables companies to model scenarios more dynamically.

The visibility gaps exposed

Another underappreciated outcome of tariff volatility has been the exposure of data gaps. Many organizations lacked detailed country-of-origin data at the material level, Schenecker said. Supplier certifications were incomplete. In some cases, tariffs were embedded in blended material costs without discrete visibility.

When trade volatility intensified, procurement teams found themselves reconstructing exposure from fragmented datasets. Schenecker pointed out that without clear origin mapping and supplier-level cost disaggregation, risk modeling becomes guesswork.

From cost leadership to risk leadership

What this shift in strategy is doing is putting procurement front and center. Procurement is moving from cost leadership to risk-adjusted value leadership, Schenecker said. Aggregation is still a tool to be used, but is no longer the default strategy.

Even if specific tariffs are reduced, the precedent of trade policy volatility remains.

“If politicians are going to use tariffs as a weapon … we should just make that assumption,” Schenecker said.

Procurement leaders are now redesigning their playbooks and integrating network design with sourcing strategy, modeling tariffs as permanent variables, and accepting that resilience carries a cost.

Executive takeaways

Editor's note: The SCM thesis Buffer or Suffer: Dynamic Multi-Echelon Inventory Optimization in Action was authored by Vi Duong and Nic Holwerda, and supervised by Dr. Eva Ponce (eponce@mit.edu). For more information on this research, please contact the thesis supervisor.

Wake-up call: 57 days of supply – still missing the mark

In today’s retail world, too much inventory is as risky as carrying too little. One U.S. grocery chain, operating a hub-and-spoke distribution model, held 57 days of supply for dry food. Inventory turnover was low, safety stocks were excessive, and service levels still varied widely.

This issue? As business expanded and SKUs multiplied, traditional inventory policies fell short of managing the network effectively. To regain control and strike a better balance between availability and capital efficiency, the company adopted dynamic Multi-Echelon Inventory Optimization (MEIO), a network-wide planning approach.

When one size does not fit all

Traditional inventory management optimizes one node (inventory-holding location) at a time. In contrast, MEIO treats the entire supply chain—from plants to hubs to spokes—as an interconnected system. The MEIO approach helps to mitigate the bullwhip effect often observed in traditional, or single echelon, optimization approaches. As explored in our capstone research, applying MEIO dynamically (regularly updating inventory policies based on changing conditions) can unlock even greater value.

This capstone analyzes the value of dynamic MEIO by running multiple planning scenarios for 61 SKUs across 31 nodes in our sponsor’s retail network. Using Coupa’s Supply Chain Guru—a platform powered by AI and advanced optimization— we modeled 18 scenarios combining six update frequencies (annual to weekly) and three service level targets (90%, 95%, 99%). Products were segmented by demand volume and variability to reflect differing inventory needs.

Why so many scenarios? Because one size does not fit at all.

Our analysis showed that high-variability products benefit from more frequent updates, which better align supply with volatile demand. On the other hand, stable, low-variability SKUs see minimal gains from frequent updates and may incur unnecessary costs. A one-size-fits-all approach misses these nuances and the opportunities they represent for the organization.

Segment. Target. Win: High service, lean inventory

Our most compelling finding? Dynamic MEIO reduced total inventory value by up to 63%—about $9.3 million annually—for just 61 SKUs. Even a conservative annual update cut working capital by 40%, freeing up cash without sacrificing service levels.

While high service targets (like 99%) typically demand more safety stock, frequent updates allow inventory to adjust in near real time. This makes high service levels more affordable—a major win for retailers focused on customer experience.

Savings were not evenly distributed. Over 50% came from hub-level inventory reductions while spokes, where demand is more predictable, contributed only 2%. This highlights upstream overstocking as a key opportunity and MEIO as the tool to unlock it.

The biggest gains came from moving inventory policies from annual to biannual updates. More frequent updates delivered diminishing returns, particularly for low-variability SKUs. This insight helps teams focus effort where it delivers the most value.

Final word: MEIO is a strategic trade-off

Dynamic MEIO isn’t a one-size-fits-all solution. It’s not plug-and-play—it involves upfront investment, setup time, and close coordination between planning and operations teams.

But for businesses facing high demand variability or targeting premium service levels, the value is clear.

MEIO is about more than cutting inventory. It represents a shift from static, reactive processes to adaptive, data-driven strategies. For supply chain leaders, it offers a powerful framework to balance service and capital, risk and agility.

Executive Takeaways

In February 2025, Supply Chain Management Review (SCMR) published my 12-part online article series on the Perfect Order. The article series—with accompanying “explainer” articles—was based on a single article written 20 years prior by Dr. Edward J. Marien that presented a (business) customer’s eight rights when ordering goods.

To recap Dr. Marien’s eight customer rights (8Rs) for logistics professionals, a customer deserves:

1. The Right Product
2. In the Right Quantity
3. From the Right Source
4. To the Right Destination
5. In the Right Condition
6. At the Right Time
7. With the Right Documentation
8. At the Right Cost

In the May/June 2025 edition of Supply Chain Management Review magazine, SCMR Editor In Chief Brian Straight (who edited my article series) authored a thought-provoking article titled “How supply chain leadership drives business growth, competitive advantage.” Brian’s article focused on how P&G’s unified supply chain strategies have helped them to grow the company and become brand leaders.

What I found so interesting about Brian’s article is how much of P&G’s supply chain strategy paralleled Dr. Marien’s foundational customer rights. Let me draw out the comments and quotes from Brian’s article and the relationship to Dr. Marien’s 8Rs.

1. The Right Product, The Right Time, and The Right Destination: “As the strategy has evolved, it has remained consistent in its approach to satisfy the customer’s desire and need for the right product when and where it was needed.”
2. The Right Time and The Right Destination: “It is the importance of being on the shelf, online, wherever and whenever people want to shop.” (P&G’s chief product supply officer.)  
3. The Right Quantity: “For beauty and personal care brands, that has meant winning consumers through superior innovation delivered at speed, and ensuring product availability.” 
4. From the Right Source: “Supply chain leaders at P&G align operations with corporate strategy to meet evolving consumer demands, leveraging advanced logistics, strategic supplier partnerships, and real-time demand forecasting.”
5. In the Right Condition: “It also means producing quality products so that customers have a superior experience when they use our brands.” (P&G’s chief product supply officer)

According to Brian’s article, P&G didn’t just launch head-on into their unified One Supply Chain strategy; it was an evolutionary process that began in the 1990s and matured through three distinct iterations. As P&G has proven: execution—from one end of its supply chain to the other—became its competitive edge in a commoditized world. If these words sound familiar, it’s because you’ve read them from me before.

Your company does not need to be as big as P&G to ensure that it is providing each and every customer their rights when they order a product from you, regardless of whether that customer is a business or that customer is a person.  As Brian’s article concludes, you need to view your supply chain as a driver of revenue.  Stop thinking about cutting costs and start thinking about value propositions. Because even if you are not, your competitors probably are.    

My Perfect Order article series can be viewed at: https://www.scmr.com/article/norman-katz-article-series-outlines-the-steps-to-achieving-the-perfect-order

Executive takeaways

For more than a decade, supply chain leaders have talked about visibility. Yet despite advances in transportation management systems (TMS), real-time tracking, and control towers, the challenge of having data is not the same as being able to act on it remains.

“Visibility without the ability to act and execute really isn’t that valuable,” explained Chad Raube, CEO of IntelliTrans, a provider of transportation management and supply chain visibility software for multimodal freight networks.

Tracking, Raube said, is just a data point at this point. Important, yes, but just one part of a larger picture. The answer to visibility in 2026 is the ability to synchronize data across systems, align planning and execution, and ensure that AI layers are built on trustworthy foundations.

The “single source of truth” problem

Many shippers today operate with five to 10 core systems such as TMS, WMS, carrier portals, rail systems, ocean platforms, or financial tools, he said.  Each contains its own version of the truth.

“What’s so interesting,” Raube said, “is hearing from shippers that … having a single source of truth is very difficult. A typical shipper might have five to 10 systems and the data might not be aligned so you don’t have the source of truth.”

The consequences show up when disruption hits. If systems don’t communicate or if timestamps conflict between platforms, teams lose critical time. “The TMS might say it’s due on Saturday and the WMS may say it’s on Sunday. And it creates a lot of confusion,” Raube said.

That delay in reconciliation can erase any opportunity to optimize cost or protect service. By the time planners sort through the discrepancies, “it might be a couple of days before they can actually find the right way forward.”

In volatile freight markets, two days is often the difference between resilience and escalation, he noted.

From buying systems to buying connections

Historically, shippers evaluated platforms based on feature depth. Increasingly, however, they are evaluating ecosystem connectivity. Raube said the switch has flipped in the last few years, with shippers now looking at “buying connections” rather than feature-rich systems as the prioritize interoperability.

Rail and truck provide a useful example of this. Raube explained that in many organizations, those modes operate in separate silos, sometimes even under different TMS providers. If a rail delay occurs, the ability to quickly evaluate truck alternatives depends on shared data structures.

“If your truck TMS isn’t connected to your rail TMS … by the time you figure out what your options are, it’s two days down the road,” he noted.

This is not simply a technical integration challenge. Ensuring data from different systems shares a common language is becoming a bottleneck for organizations. The standard bill of lading provides an example of this. The reality is that there is no standard bill of lading.

“No two documents are the same,” Raube said. “Even a bill of lading is different. Nobody can even agree on what a bill of lading should look like.”

Without harmonized identifiers and formats across documents and systems, even the best analytics tools struggle, he noted.

AI is not a substitute for data quality

According to one venture capital estimate, there could be 1.4 million artificial intelligence applications by 2030. But that quantity of solutions doesn’t necessarily translate to better data. AI can accelerate decision-making, but it also accelerates flaws.

“If the data’s not good, the AI on top of this is going to lead to bad decisions,” Raube said.  

Embedding AI directly into freight workflows without resolving the data inconsistencies risks making errors faster, not smarter, he added. “You cannot afford to be inaccurate,” Raube said. “While AI is very powerful, we need to make sure that whatever we apply maintains a 100% accuracy.”

Beyond dashboards: Acting on visibility

Another theme emerging across shippers is the need to move beyond dashboards. For years, supply chain visibility tools focused on tracking and reporting. Today, leaders want orchestration.

That requires more than visualization, Raube said, noting that it requires real-time API connectivity, harmonized identifiers across systems, embedded analytics that surface exceptions immediately, and clear workflows that define who acts and how.

Without those components, visibility remains descriptive rather than prescriptive.

The path forward: alignment before intelligence

Raube said that alignment needs to be a priority for everyone. Shippers, for instance, need shared data across modes, trusted identifiers, an API-driven ecosystem that prioritizes connectivity, and governance frameworks that ensure AI outputs are auditable. Without some of these basics in place, the ability to fully take advantage of AI will be muted.

While visibility has been a constant theme for years, the belief that AI alone will provide it is misguided. Raube said that synchronization across modes, data platforms, and partners will be what ultimately delivers the visibility, the accuracy and the resilience the supply chain is looking for.

Executive takeaways

A term often used in discussions to improve supply chain performance is ‘balancing demand and supply.’  But what does that mean in real-world, practical applications?

In short, balancing demand and supply means doing what it takes to have demand = supply. In a perfect world there is a demand plan (expressed as units shipped for a given item or SKU on a given day/week/month). A supply plan is generated to match the demand plan; the work is ensuring materials, production and distribution plans are sufficient to match demand plan.

 

Good supply chain planning balances demand and supply in an efficient cost-to-serve manner.  Conversely, poor planning leads to operational heroics to meet orders. These heroics are likely familiar to planners because they manifest as overtime, changeovers/poor OEE, expedited transportation, increased inventory and lead to poor bottom-line results for the enterprise.

Even if perfectly planned (and avoiding the aforementioned operational heroics to meet incoming orders), a balanced plan can be elusive especially at a SKU/item/part level. Varying lead times and production resource capacities can create gaps. Broadly, there are two scenarios a planner needs to work through when demand is not equal to supply:

1. Demand is greater than supply. Safety inventory is typically the first buffer to close gaps when the demand plan is greater than supply (and by definition, safety stock exists to be used—if it’s not being used at any point, then safety levels are too high…) If safety inventory is insufficient, then the work to rebalance would include (1) checking ability to transfer inventory within the network, (2) to produce more, (3) to substitute an acceptable equivalent product (e.g. same UPC),  (4) to check the possibility of shifting demand to similar products available or that can be produced (e.g. different size or different flavor/scent) or (5) if demand is being stimulated by marketing/promotions, look to adjust (delay or refocus to available products.
2. Demand is less than supply. In this case, the work to rebalance is either (1) reduce supply—cut or delay production or deployments or (2) stimulate demand—work with sales/marketing, especially if production has short or finite shelf-life and would lead to scrapping or sales at significantly reduced realization.

Why is balancing demand and supply important? While a demand/supply imbalance may be seen first at a supply planning level, it’s actually a true multi-functional business challenge. Financial plans typically tie to demand outlook. Implications of demand being unequal to supply can hit financial top and bottom lines. It’s intuitive to correlate lower demand with poorer financial outcomes. But if the gap is supply related, the impact can also be lower revenue (top line) and profit (bottom line.) For example, marketing spend is typically justified by an expectation to generate a specific demand amount. With supply insufficient to demand; companies could spend marketing dollars without achieving the incremental sales goal, impacting profit. Similarly, if demand is insufficient vs. supply plans, financial risks include product scrapping or lower realization that can negatively impact financials. But risks are not limited to immediate financial impact; supply being less than demand can create customer ill-will, imperiling future business opportunities, for example customer support of new product introductions in future.

Related:

What It Really Means: Bringing the outside in

What it really means: Democratizing the data

What it really means: Supply chain control towers

Benefits to balancing demand and supply: Competing in the market is hard, so it’s important to give your company the best chance of succeeding, which includes avoiding under or over supply vs. expected sales. The benefits of a balanced demand and supply plan show up in areas like:

• increased sales revenue and profit
• reduced expediting
• increased operational stability
• increased marketing and promotion ROI
• reduced short term plan changes (up or down)
• reduced sales at lower realization (discounting)
• reduced scrapping  

Watchouts: Unfortunately, there can be many intended or unintended barriers to balancing demand and supply, including:

having a ‘one-size-fits-all’ solution when demand outstrips supply or supply outstrips demand.  Companies should use portfolio segmentation to determine strategies and thresholds by product/customer to guide actions (e.g. some products require service at 100%, others may not merit  the additional expense of meeting every unit of demand).

• lack of multi-functional collaboration, especially communications about activity to stimulate demand not yet reflected in plans
• insufficient attention to plan quality beyond short-term (unintentionally consuming lead times that may be needed)
• not planning using demonstrated capabilities (can create insufficient supply)
• not having inventory to absorb normal short-term variation (timing or product mix)

How to balance demand and supply?

The good news: the best method to balance demand and supply is not new. This method is strong multi-functional collaboration in the sales & operations planning cycle (S&OP). A healthy S&OP includes using marketing and sales inputs to help create high-quality demand plans, and using the demand plan to generate a matching supply plan based upon demonstrated capabilities (capacities, production rates, lead-times, etc.) and an inventory plan that can absorb short-term variability (e.g. timing of demand from week to week or shifts across product lines).  

About the authors

Andrew Byer is a former P&G Supply Chain Leader. Mike Dobslaw leads EY’s Supply Chain Planning Practice.  To learn more about how EY and P&G team to support supply chain transformations please write michael.dobslaw@ey.com

Executive takeaways

In mid to late 2025, cargo theft was among the top issues facing the freight market. Then English-language proficiency enforcement took the stage, followed by non-domiciled CDLs. There was the continued freight recession chatter.

None of those have gone away, but while the news cycle continues to churn out dramatic headlines, the underlying risk environment for carriers, brokers and their shipper partners, remains as fluid as ever. For freight leaders, the challenge is that the traditional signals around rates, capacity, and enforcement that once guided the market are becoming harder to interpret amid overlapping risks.

“I would say you don’t hear as much about it now,” said Kendra Phillips, VP of global transportation management at Ryder, told Supply Chain Management Review about cargo theft at the Manifest conference earlier this year. “The noise has quieted down, but it’s obviously still a very big concern in the industry.”

It’s just that there are so many risks today that no single issue is able to stay in the headlines for too long.

Cargo theft: Fewer headlines, more sophistication

Phillips said she believes cargo theft is actually worse than it was five or ten years ago, but the nature of the threat has evolved. Traditional theft at rest is now joined by increasingly organized identity fraud.

“You even see it in rail,” she noted. “People are asking, ‘Hey, I want to put my product underneath [on double stack trains]. I don’t want it on the top.’”

More concerning, though, are the fake carriers that brokers continue to battle.

“All the fake carriers that come up and they act as a credible carrier and get into your network for a couple of months … and then they take a load and they disappear,” she said.

She referenced what some describe as a “carrier mafia,” creating shell companies that perform legitimately long enough to build trust before stealing freight. Scott Cornell, vice president of transportation risk and strategy at LogistIQ Insurance Solutions, recently joined the Talking Supply Chain podcast to talk about cargo theft, noting that cargo theft incidents increased roughly 93% between 2021 and 2024, while strategic cargo theft—fraudulent schemes involving brokers, carriers, and falsified documentation—surged more than 1,400%.

For shippers used to vetting a carrier one time, the calculus has changed, Phillips said.

Vetting isn’t static

Phillips outlined a layered approach to carrier risk management—financial health monitoring, safety records, compliance flags and capacity validation.

“If you say that you only have 10 trucks in your fleet, are you accepting 12 loads? You can’t be,” she said, adding that continuous monitoring is critical. “It’s one thing to vet a carrier once and use them for three years. You need to be going down many layers … and make sure you’re getting continuous updates.”

She emphasized that shippers should be asking 3PLs hard questions such as:

• How deep does your vetting go?
• How frequently are carrier updates reviewed?
• How do you test new carriers before placing shipper freight?

CDL enforcement and English proficiency: Flashpoint or structural shift?

One of the most discussed freight topics in recent months has been enforcement around non-domiciled CDLs and English-language proficiency. Phillips confirmed there was a real impact late last year.

“As you went into Q4, there was a very sharp tightening,” she said.

Phillips, though, believes the combination of ramped-up enforcement and immigration pressures added to rate increases.

“We had a lot of carriers turning down loads … if they had to go into ICE-heavy areas,” she said.  

Taken together, all the factors have resulted in measurable capacity tightening. However, Phillips noted that enforcement intensity appears to have eased. “It feels like the enforcement has lightened up … but that’s a feeling, that’s not fact,” she observed.

From a shipper liability perspective, the issue remains serious. “They should be absolutely 110% concerned,” Phillips said.

In today’s nuclear verdict environment, any accident involving language proficiency or licensing questions could cascade upstream, resulting in brokers and shippers paying a heavier price.

AI in carrier vetting: Helpful, but not enough

With rising fraud and compliance risk, many logistics providers are leaning heavily into AI for carrier vetting. Phillips believes AI plays an important role, but should not replace human oversight.

“We let AI do a lot of it … but at the end it has to hand it to the human for the final, [sign off]” she said.

According to Phillips, that human touch is important as a “human’s going to suss out [if something] feels wrong or off?”

That human layer carries cost, but Phillips sees it as theft prevention insurance.

Freight rates: The inflection point approaching?

Phillips touched on a number of other issues facing the industry, including the recent rise in freight rates.

Phillips sees classic supply-demand dynamics playing out. “If you look year over year, freight volumes are down … about 6%,” she said. “Retail sales were flat … after products cost a lot more now than they did a year ago.”

Flat dollar sales, adjusted for inflation, imply declining physical volume, she noted. At the same time, capacity is exiting the market for many of the reasons mentioned previously. Capacity appears to be leaving the market faster than freight demand slows, creating an imbalance that could trigger an inflection point this year, Phillips said.

Emissions: Regulation may shift, shipper expectations haven’t

One approach the Trump administration has taken is to ease and even eliminate some of the more restrictive emissions regulations put in by previous administrations. But, while the requirements may be easing, Phillips doesn’t see the pressure from shippers easing.

“So many of our shippers are still very environmentally conscious,” she said, noting that reporting expectations remain regardless of the regulatory shifts. “They still want to know … what’s the impact I’m having on the environment?”

The bigger picture

If there is a common thread in Phillips’ perspective, it is that freight risk today is multi-layered. Cargo theft is more sophisticated. Fraudulent carriers are harder to detect. CDL enforcement can tighten capacity quickly. Freight volumes remain soft. Capacity is exiting faster than demand warrants.

The result is that there is more risk today than ever before. And as Phillips makes clear, technology alone won’t solve it.

Executive takeaways

Every company believes it is unique. Its operation is unique. Its challenges are unique. And as a result, when it comes to installing warehouse automation, it needs a unique solution.

But are more tailored solutions the best approach to every challenge? Probably not. In some cases, the more tailored the solution, the more expensive it may be turn out to be in the long term, eliminating the exact reason to install automation in the first place—to streamline operations and reduce costs.

Romain Moulin, CEO and co-founder of Exotec, told Supply Chain Management Review that the assumption that a customized automation solution is the best approach can hamstring organizations for years to come.

That level of customization seems like the ideal approach, but in practice, it often undermines cost, reliability, and scalability, he said.

“We are in an industry which has been used to really offering a bespoke solution—designing a machine and software around the customer’s needs,” Moulin said. “Which is very good for suiting the customer needs, but which is very bad for time, cost and reliability.”

The hidden cost of custom automation

During his on-stage presentation, Moulin compared bespoke warehouse automation to first-of-a-kind megaprojects that are complex systems that inevitably exceed budgets and timelines.

Inside the four walls of a facility, the risk shows up in three ways:

1. Performance uncertainty. Individual modules may be engineered precisely, but predicting overall warehouse performance becomes difficult.
2. Operational fragility. Highly customized software depends on specific individuals who may no longer be available when problems arise.
3. Slow ramp-up. Complex systems take longer to stabilize, delaying ROI.

“The more assumptions you make, the less resilient your warehouse is” Moulin said.

If every warehouse in a distribution network is uniquely engineered, balancing throughput or shifting inventory becomes nearly impossible, he argued.

Standardized doesn’t mean one-size-fits-all

But Moulin is not advocating for cookie-cutter solutions. In fact, he pushes back when the idea of building a one-size-fits-all solution is brought up. Instead, he advocates for a Lego-block type approach. Lego is known for creating standard blocks and letting kids’ imaginations go wildly with the designs. That is the approach that Exotec is taking.

“All of our solution are different for all the customers, but they are made of the same building blocks,” he said. “It’s really the Lego … which is reassembled differently”

In essence, the hardware modules, the software layers, and the robotic elements are standardized. Their configuration is not.

“You should not say you have one solution for your customer … but you listen to your customer and you reassemble your bricks together [to meet their needs],” Moulin said.

Composability enables replication and replication is what drives speed and cost control, he added.

Designing for uncertainty, not forecasts

Moulin also warned against over-designing around forecasts. Warehouses are typically built on detailed projections: B2B vs. B2C mix, order profiles, lines per order, or cartonization assumptions. But those ratios can change. Moulin offered the example of a cross-channel warehouse that shifted from 80% B2B and 20% B2C to the inverse within three years.

If the organization had installed a static automation solution based on the original mix, the cost savings achieved in those first three years may evaporate as it retools for the new product mix.

“If your warehouse cannot cope with that, you are really in danger,” Moulin said.

Rather than hard-coding assumptions into workflows, he advocates designing systems that rely on optimization algorithms rather than preconceptions.

“We won’t program the solution to the problem,” he explained. “We will use mathematics to find the best solution at any given moment.”

Replication as strategy

In his presentation, Moulin described how this philosophy scaled across a major sports retailer’s network. Rather than launching sequential four-year bespoke projects, the company defined a template warehouse with variable robot fleets and inventory sizing. The result: faster deployment and reduced inventory levels.

From a strategic standpoint, replication also reduces integration risk, he argues. Moulin says projects are evaluated by how much custom software labor is quoted.

“It tells me exactly the risk associated with that project,” he said.

The smaller the bespoke layer, the lower the ramp-up risk and the easier upgrades become Moulin noted.

Starting with the problem, not the solution

The right approach to automation in Moulin’s view is similar that being advocated for the adoption of artificial intelligence. Don’t start with a solution and try to find the problem, start with the problem and find the right solutions.

“Tell us about your logistics and we’ll work together to find what the right approach,” he said.

Customers often come with the idea that then need to install ‘X’ solution, but that isn’t always the right approach. The right solution is the one that solves the pain point and is scalable and flexible as needs change.

That may be a customized solution, or it may be a new configuration of existing solutions. Moulin doesn’t argue against customization, but rather for finding the right solution that can adapt with your business as you grow.

Artificial intelligence has dominated supply chain conversations for several years now. Yet for all the enthusiasm around generative AI and agentic systems, organizations that have moved passed pilot projects remain small in number.

“I would argue that most enterprises are not yet adopting AI and going into production,” said Aadil Kazmi, head of AI at Infios, told Supply Chain Management Review during a recent discussion on the state of AI deployment at the Manifest conference.

While generative AI remains “top of mind for executives,” he noted, scaling it into day-to-day operational decision-making is far more complex than launching a proof of concept.

The pilot trap

Kazmi described two approaches for understanding AI adoption. Within Infios’ own customer base, some deployments have progressed rapidly from pilot to production in “a matter of two quarters.” But across the broader ecosystem, progress has been slower. The disconnect, he suggested, often begins with how organizations frame the problem they are trying to solve.

“A couple of weeks ago, one of our own customers came to us and said, ‘Hey, we’re looking to optimize how we build loads.’ And they kept on approaching it as a gen AI problem,” he said. Instead, Infios redirected them to a classical machine learning optimizer already embedded in its platform.

The example, Kazmi explained, is reflective of the approach many are taking when it comes to AI. Find a problem to solve rather than attacking problems and identifying the correct solution. Leaders feel pressure to “do AI,” often before validating whether AI models are even the right solution.

Kazmi offered a framework grounded in principles. “What changed three years ago?” he asked. Machine learning, natural language processing and OCR have existed for years, but the shift occurred when large language models could finally understand unstructured data and communication.

“If we agree that the fundamental change three years ago … was machines can now understand unstructured data using LLMs, then the framework should be … to look at workflows that already depend on unstructured text and data as the key blocker for automation,” Kazmi said.

In other words, AI agents are best applied where human reasoning over emails, voice calls, and free-form communication has historically prevented automation, not where traditional optimization or RPA already suffice.

Embedded agents vs. bolt-ons

Another design choice that shapes success is where agents live. Kazmi described two philosophical approaches. The first is vertical—automating a specific workflow such as driver check calls. The second is horizontal—connecting multiple systems (ERP, TMS, WMS) to create cross-functional automation.

Equally important is whether AI is embedded into existing software or layered on top as a standalone interface. “When we deploy our order entry agent … it natively lives inside [the system],” he explained. Users forward emails to the agent without leaving their core platform. By contrast, a bolt-on interface forces users to toggle between systems, increasing friction and limiting adoption.

Beyond chatbots: From data to action

Kazmi cautioned against stopping at conversational AI, though. “Chatbots are fantastic,” he said, particularly for querying reports, but frontline supply chain teams do more than retrieve information—they execute.

“Their core workflow isn’t just querying data, it’s taking actions,” he said.

To generate measurable ROI, AI agents must interpret data but also perform tasks such as updating order status, canceling shipments, or scheduling appointments. That requires integration into transaction systems and access to APIs, not just access to reports.

Guardrails, governance and autonomy

As AI agents move closer to operational execution, governance becomes critical.

Kazmi acknowledged the risks of agents deviating from business rules but said that preventing such failures is a matter of architectural discipline.

He introduced the concept of L1 versus L2 autonomy. “L1 being specifically gated to the task at hand … versus L2, which is true decision autonomy,” he explained.

Infios recommends starting with L1 task automation, where ROI can be substantial. “Some of our customers are seeing ROI of upwards of 60% with just L1 automation,” Kazmi said.  

Decision autonomy can follow later once governance, confidence thresholds and human oversight are mature enough to handle it. He also emphasized reinforcement learning through human feedback to prevent agents from institutionalizing exceptions.

Ultimately, he advised managing AI agents much like human employees: monitor performance, set guardrails, and manage risk tolerance.

Purposeful innovation at scale

One of the most difficult balancing acts for software providers is managing demand.

“We have a ton of customers coming to us with every single sort of problem,” Kazma acknowledged. But rather than building bespoke agents for every request, Infios maps workflows across its customer base to identify scalable use cases, in a sort of Venn diagram approach: identify common workflows across hundreds of customers, then prioritize those with the greatest repeatable value.

This “purposeful” strategy shortens deployment cycles over time, Kazmi said. Early pilots create learnings that compound across subsequent implementations.

Measuring ROI: Follow the workflow economics

Post-deployment, the question shifts from excitement to economics. ROI can vary, Kami pointed out, with some use cases generating immediate returns while others take months due to required training and change management.

To measure value, he recommends starting with workflow-level cost analysis.

“If you’re looking at a workflow to automate, start with what is the current process? How many humans are touching this … what is their average salary or hourly rate?” Kazmi suggests. From there, calculate cost per execution and compare it to the AI-driven alternative.

While time savings are often cited, Kazmi argues that cost-to-serve is the clearer metric, particularly in high-volume operational environments.

The human dimension

Finally, scaling AI requires more than technical integration. Infios provides pre-built templates for common supply chain workflows, reducing the burden on non-technical users. But it also pairs deployments with professional services support and change management.

AI agents may automate tasks, but humans still define guardrails, validate outputs and adapt processes, Kazmi noted. As with previous waves of automation, upskilling, not replacement, remains central to sustainable impact.

From hype to operational discipline

If there is a common thread in Kazmi’s perspective, it is this: AI success in supply chain is less about model sophistication and more about workflow clarity. Select the right problems. Embed agents where work already happens. Start with constrained autonomy. Measure ROI at the task level. And treat governance as a design principle, not an afterthought.

The industry may still be in the pilot-heavy phase. But the path to production is becoming clearer, one workflow at a time.

Executive takeaways

The life sciences, and in particular, the pharmaceutical supply chain, are far more complex than other products and industries. Pharmaceuticals and other medical products require unique compliance with FDA regulations and control. Ingredients are often scarce and difficult to procure.  Batch sizes can be in the millions all the way down to a batch size of one. Products often have to be temperature-controlled. This complexity sets this industry’s supply chains apart from most others and presents unique challenges for supply chain planning and operations.

The global nature of the pharma industry

The pharma supply chain is global. Building-block chemicals that are used for common off-the-shelf drugs such as acetaminophen and antibiotics are manufactured almost entirely in China and India. Other pharmaceuticals use unique and scarce ingredients that are sourced worldwide from plants and animals.

From individualized treatments to massive production of building-block chemicals, the manufacturing and shipment of these products to consumer markets is global, often originating in China and India and delivered worldwide. The risks of these long global supply chains and America’s complete dependence on imports for drugs such as antibiotics were fully exposed during the pandemic. This resulted in new regulations, massive funding opportunities for U.S. expansion, and the reshoring of some production.

Manufacturing, shipping, and logistics

In addition to long global supply chains, many pharma products incorporate unstable ingredients that must be kept in a temperature-controlled environment. This poses major challenges when sourcing active pharmaceutical ingredients, referred to as APIs, which are the core, biologically active components responsible for a drug’s intended therapeutic effect. They are part of the medicine that treats the disease, relieves pain, or prevents symptoms, and they are often rare and scarce.

Related podcast

 

Frictionless Supply Chain: The pharma supply chain

In this episode of the Frictionless Supply Chain, Beth Hayes, a consultant to pharmaceutical companies, joins host Rosemary Coates, executive director of the Reshoring Institute, to discuss the complexity of pharmaceutical supply chains. Pharmaceuticals and other medical products require unique compliance with FDA regulations and control. This complexity sets this industry's supply chains apart from most others. 

Listen today

 

Manufacturing of a finished product is typically outsourced if the drug is made in mass quantities, as is the case with off-the-shelf products such as Tylenol and Excedrin, which are made overseas. This introduces issues of oversight and regulatory adherence. Stringent requirements like the U.S. FDA Drug Supply Chain Security Act (DSCSA) require complex, full-batch traceability, making it difficult to quickly switch manufacturers or suppliers.

Other drugs, particularly personalized drug treatments, are made in small lots or even in a lot of one, making production of these life-saving products difficult to plan. Manufacturing must be precise, often includes controlled substances and sterilization processes, and all of this is carefully controlled by government regulations.

Raw materials as well as finished products may require carefully monitored cold-chain shipping and protection from contamination. Roughly 70% of the top-selling pharma products require temperature-controlled logistics, with about $35 billion lost annually due to temperature-controlled failures and improper handling.

Geopolitical and cybersecurity risks

We can easily imagine our dependence on China and India for drugs such as antibiotics. Geopolitics play an important role in securing the supply. Concentration of drug manufacturing in one country or region results in vulnerabilities caused by politics and policies, potentially threatening the supply of drugs to the U.S. or other countries which could be devastating to populations.

Cybersecurity is yet another risk in the global supply of drugs. Contamination of products or disruption of distribution through cyberattacks are possibilities.  While cybersecurity is a major concern in all industries, the pharmaceutical industry is particularly vulnerable.

These are not everyday supply chain challenges

Pharmaceutical industry supply chains and logistics are difficult, requiring skilled and experienced supply chain professionals to source, plan, produce, package, and distribute drugs. The overlay of government regulations adds a unique and additional layer of complexity. Geopolitics plays a critical intervening role. Hats off to our supply chain colleagues working in this challenging industry.

Executive takeaways

Operations and supply chain roles increasingly demand graduates who can work inside messy systems. New hires are expected to analyze unstable processes, interpret incomplete data, and engage frontline stakeholders in real time. Traditional lecture-heavy courses and textbook cases teach concepts well, but they often fall short in developing applied capabilities, such as lean waste identification, current-state mapping, and implementation-focused problem-solving. In a recent operations management course, students who worked inside a live operating facility reported markedly stronger applied learning outcomes than peers who completed traditional textbook case projects. As shown in Table 1, students participating in industry-based projects reported meaningfully stronger applied capabilities than those completing textbook cases. Gains were largest in applying operations methods and lean thinking, with improvements of roughly 10% to 14%. Differences in higher-level conceptual understanding, such as operations strategy, were present but more modest.

Table 1:  Learning domain in different project formats

Note. Mean post-course scores (1–5 scale) based on the author’s course data.

Immersive learning offers a way to close this gap. Rather than simulating reality, immersive projects place students directly inside operating systems where ambiguity and judgment are unavoidable. Drawing on an immersive engagement model implemented in an undergraduate operations and supply chain management (OSCM) course, this article demonstrates that repeated engagement with a live facility function not only serves as teaching but also as early-stage training for real-world supply chain work.

Why immersive projects matter

Applied capabilities develop most strongly when learners work in authentic settings and repeatedly test assumptions against reality (Kolb, 2015). To examine what this looks like in practice, a 3000-level OSCM course at a Midwestern university employed two project formats in parallel. Two sections (n=70) completed structured textbook case projects, while one section (n=29) completed an industry-based immersive project with a regional material recovery facility (MRF). The study is consistent with prior research in operations and supply chain education that compares case-based and experiential formats (Heriot et al., 2008; Kang et al., 2010). All sections shared the same instructor, lectures, and assessments.

From classroom project to immersive engagement

The immersive project was intentionally designed to move beyond a traditional classroom assignment and function as a structured, real-system engagement. The course partnered with a privately operated material recovery facility, which provided a complex, high-variability operating environment, making it well-suited to applied learning in operations and supply chain management. Students were positioned as novice analysts rather than problem solvers. They were not asked to redesign processes or implement solutions, but to observe work as it actually occurred, verify assumptions with frontline employees and line managers, and document current-state processes. This design choice shifted the project from a one-time course exercise to an immersive experience that mirrors how early-career professionals learn to understand and diagnose real operational systems.

Inside the four-visit immersive model

The immersive project embedded students directly in a high-variability, sustainability-oriented operating environment. Students were organized into four teams, each assigned to a specific production line for the duration of the project. The engagement followed a repeatable, worksheet-driven cycle across four site visits.

Visit 1 – Orientation and scoping. Students completed safety training, toured the facility, and observed all production lines, clarifying project scope and initial task constraints with line managers.

Visit 2 – First-pass mapping and data capture. Teams observed their assigned line, documented task steps, equipment use, handoffs, and visible wastes, and sketched initial current-state maps using structured worksheets.

Visit 3 – Refinement and validation. Students returned with revised maps, conducted targeted observations, and validated assumptions with line managers and operators. Each visit ended with a simple check: “Is this what really happens on your line?”

Visit 4 – Final validation and synthesis. Teams reviewed near-final maps with facility staff, corrected inaccuracies, and synthesized findings into a final report and presentation.

At the end of the semester, the community partner attended in-class presentations, heard student recommendations, and provided feedback and clarifications. This closed the loop between student learning and organizational value.

What changed in student learning?

Learning outcomes were assessed using voluntary pre- and post-course surveys aligned with course objectives. Four domains were measured on a five-point scale: operations strategy understanding, process understanding, lean thinking, and application of methods.

Pre-course results indicated similar baseline levels across sections, suggesting comparable starting points. Post-course results, however, revealed clear differences by project format. Students in the industry-based immersive project reported higher mean scores across all four domains than students in textbook-based projects. The largest differences appeared in applied domains.

Figure 1: Post-course learning outcomes by project format (mean scores)

(Photo: Author)

The largest gains appeared in applied domains such as lean thinking and application of methods, while improvements in higher-level conceptual understanding were smaller, a pattern commonly observed in experiential operations and supply chain education (Al-Shammari, 2022; Yazici, 2020). Regression analyses controlling for year of study, prior operations coursework, work experience, and familiarity with operations confirmed that project format remained a significant predictor of post-course outcomes across domains.

In practical terms, students who worked inside a real system felt more confident identifying waste, applying lean tools, and interpreting operational tradeoffs than peers who worked exclusively with cases. From an employer perspective, these gains translate into faster onboarding and earlier contribution in improvement and operations roles.

What the numbers mean in practice

The pattern aligns closely with experiential learning theory. The immersive project exposed students to real process variability and incomplete data; repeated cycles of observation, mapping, verification, and revision; direct interaction with line managers and operators; and responsibility for presenting feasible improvement ideas. The importance of these conditions has been consistently noted in experiential and project-based operations education (Heriot et al., 2008; Miyaoka et al., 2018). The stronger gains in lean thinking and application of methods suggest that students learned not only what lean is, but also how to use it in real systems.

Why this functions as training, not just teaching

The MRF presented a highly variable context characterized by contamination, equipment downtime, and staffing constraints. Students confronted the same ambiguity and noise that characterize real operations. Further, the four-visit structure, paired with worksheets and explicit expectations for verification and revision, mirrored the iterative nature of improvement and consulting work. Students worked alongside line managers and employees to develop current-state maps and waste analyses, and then presented the findings to those stakeholders. Additionally, these dynamic parallels early career roles, where credibility must be earned through engagement rather than authority. Immersion most strongly affected applied capabilities, while conceptual strategy knowledge improved more modestly, consistent with prior findings that case-based approaches support conceptual understanding, while immersive projects develop judgment and application skills (Kang et al., 2010; Yazici, 2020).

Implications for universities and industry partners

The immersive training is beneficial for both educators and companies (Table 2). For educators, immersive, industry-embedded projects demonstrate that applied capabilities can be developed and assessed systematically through structured, repeated engagement rather than one-off experiences. For companies, the same model provides a low-risk way to surface operational issues, gain current-state insight, and engage with potential hires without committing to a full consulting or improvement program.

Table 2: Value of Immersive, Industry-Embedded Projects

Final wrap up

Immersive learning prepares students for the reality they will face on day one: real systems, real people, and real constraints. Evidence from this engagement indicates that students who worked in a live operating environment developed stronger applied capabilities than peers who relied on textbook cases. For supply chain leaders and educators alike, immersive projects offer a practical means of aligning education with workforce needs. When carefully designed, they function not only as teaching tools but also as early-stage training for the next generation of supply chain professionals.

About the author

Corrine Chen is an educator, researcher, and former industry executive with over a decade of hands-on experience in supply chain management, procurement, and innovation. She teaches supply chain management courses at the University of Nebraska Omaha. Corrine’s work bridges academia and practice through published research, applied projects, and a commitment to empowering the next generation of supply chain professionals. She can be reached at peonyhill@yahoo.ca.

References

Al-Shammari, M. M. (2022). An exploratory study of experiential learning in teaching a supply chain management course in an emerging market economy. Journal of International Education in Business, 15(2), 184–201. https://doi.org/10.1108/JIEB-09-2020-0074

Heriot, K. C., Cook, R., Jones, R. C., & Simpson, L. (2008). The use of student consulting projects as an active learning pedagogy: A case study in a production/operations management course. Decision Sciences Journal of Innovative Education, 6(2), 463–481. https://doi.org/10.1111/j.1540-4609.2008.00186.x

Kang, R., Yang, J., & Wei, J. (2010). Engaging undergraduate business students in experiential learning through a required term project in an operations management course. Educational Research, 1(4), 99–111.

Kolb, D. A. (2015). Experiential learning: Experience as the source of learning and development (2nd ed.). Pearson Education.

Miyaoka, J., Ozsen, L., Zhao, Y., & Cholette, S. (2018). Experiential undergraduate operations management course engages students. Journal of Supply Chain and Operations Management, 16(3), 219–247.

Yazici, H. J. (2020). Project-based learning for teaching business analytics. Decision Sciences Journal of Innovative Education, 18(4), 589–611. https://doi.org/10.1111/dsji.12219

Executive takeaways

Gartner predicts that 40% of enterprise applications will include AI agents by the end of 2026. That’s up from less than 5% in 2025. By 2028, they expect 15% of day-to-day work decisions to be made autonomously. A number that, today, is essentially zero.

Those statistics should change the way supply chain leaders think about hiring. Today, the conversation centers on whether AI can be trusted to make decisions. But that conversation is winding down. Tomorrow’s big question is, why are humans still manually executing work that AI can handle on its own? And the reality is, if AI is doing more of the executing, the talent you hire needs to look very different. The playbook most companies use, recruiting for process knowledge, tool proficiency, and execution speed, was built for a task-based world. That world has a shelf life.

The world of autonomous execution means we need a new definition of “routine.” It used to mean simple, repetitive, low-stakes. In 2026, multi-agent systems will coordinate complex workflows end to end, across entire supply chains, and the tasks they handle won’t feel small. They’ll feel like work that, until recently, required experience and judgment.

AI agents are already investigating delayed shipments around the clock, coordinating with carriers, and updating stakeholders. They’re reading shipping documents, extracting structured data, and creating shipment records. Managing dock appointments across email and portals and phone. They are requesting proof of delivery, running follow-ups, and validating documentation. They are responding to customer inquiries in real time, processing claims, identifying detention and demurrage disputes, assembling evidence, and managing resolution workflows.

These agents don’t operate in isolation, either. One extracts a document, another monitors the shipment it created, a third adjusts the dock appointment based on an updated ETA, and a fourth requests the POD when the shipment arrives. AI handles decisions where the variables are complete and the parameters are clear.

What stays human is everything that requires social capital

The World Economic Forum’s Future of Jobs Report found that 39% of existing skills will be transformed or become obsolete by 2030. But the skills rising fastest aren’t technical ones. They are those that require creative thinking, resilience, and analytical thinking. For supply chain, at least three human advantages stand apart and are superior to AI.

The first is reading between the lines. It’s understanding what a supplier isn’t saying. It’s detecting when a carrier is being evasive versus cautious. It’s picking up on signals that precede problems before they surface in data. AI can analyze communication patterns all day. A human knows that when your supplier says, “We’re monitoring the situation,” what they mean is “Start looking for alternatives.” That distinction doesn’t live in a dataset.

The second is building trust over time. It’s when a carrier prioritizes your loads during peak season because your relationship manager helped them through a capacity crunch last quarter, not because your AI submitted the most optimized request. It’s when candid conversations identify issues before they escalate. As a result, you get prioritized when things are tight. You hear about problems early, from people who pick up the phone because they know you, not because a system flagged it. That kind of capital accumulates slowly, and only between people.

The third is navigating corporate politics. It’s getting cross-functional alignment inside an organization where not everyone agrees or even likes each other is a distinctly human function. It’s knowing that finance still doesn’t trust Ops because of a project that went sideways two years ago. It’s understanding which stakeholders need to feel heard before they’ll sign off, and which ones just need the numbers. AI can propose an optimal solution. A human knows why the CFO will reject it and how to reframe it so it gets approved.

A lot of current supply chain roles amount to rubber-stamping what AI has already recommended. People review decisions they don’t have time to evaluate. Approve things they didn’t meaningfully assess. Everyone quietly knows this, but org charts haven’t caught up.

Supply chain leaders need to rethink three things.

1. Roles. Stop hiring for tactical execution and start hiring for relationship-building, critical thinking, and organizational savvy.
2. Training. Skills development should center on negotiation, emotional intelligence, and stakeholder management, not tools, which can be taught
3. Decision rights. Be explicit about what AI decides on its own versus what genuinely requires a human. Stop pretending people are reviewing decisions when they’re really clicking “approve.”

The technology is ready. Most organizations haven’t redesigned the work to match.

We’ve moved past debating whether AI can execute supply chain decisions. It can, and it will. The question worth spending time on now is what your people should be great at once the execution layer is handled. Most companies haven’t answered that yet. Some haven’t started asking. But the org charts we’re hiring into today will look very different in two years, and the roles that survive will be the ones where a human is doing something AI genuinely can’t.

About the author

Amanda Dyson is VP of marketing at FourKites. She has more than 20 years of experience leading go-to-market strategy, demand generation, and brand growth for enterprise AI, supply chain, and ERP software companies, and has led regional and global marketing teams across a range of high-growth SaaS organizations.

Executive takeaways

Over the past decade, global business has learned to operate under the assumption that disruption is normal. Tariffs come and go, geopolitical tensions flare. What has changed is not the frequency of disruption, but the realization that volatility is now a permanent feature of the system.

Hence, while renewed geopolitical and trade tensions between the U.S. and key partners dominated the Davos news cycle last month, the mood among supply chain leaders was calm. The fact is, most large, global supply networks are already designed to operate under this level of uncertainty, because there was no other choice.

Why supply chains are holding up

Today’s supply chains are built for survival. Companies have built operational muscle through repeated exposure to tariff threats and trade wars. They’ve become accustomed to pulling shipments forward, rerouting inventory across regions, adjusting sourcing strategies, and relying on multi-node networks designed to dampen shocks.

Further escalation of trade tensions is unlikely to result in empty shelves or stalled production lines. From an availability perspective, the system is functioning exactly as intended. But it’s not cheap.

This shift didn’t start with the current U.S. administration, or even the first Trump presidency. Long before trade wars dominated headlines, supply chains were already moving toward regionalization and reduced dependence on single countries. Global supply chains are far more robust than they used to be, which is reflected not only in operational performance, but also in the muted reaction from investors to policy noise.

Trade policy matters, but is no longer the point of failure

In the weeks following Davos, attention turned to the delay to ratifying the EU-U.S. trade deal. But while lower-friction, lower-cost trade is always preferable, this kind of policy freeze won’t upend global trade flows. The goods will keep moving while politics catches up.

At the same time, governments are actively looking to diversify trade agreements in response to growing instability—the EU’s trade deal with India, Canadian and British visits to China. Even though the EU missed out a deal with several nations in South American trade bloc Mercosur, efforts to find ways around U.S. pressure will continue.

These moves point to a more fragmented and reactive trade environment, one that disperses uncertainty rather than eliminating it. For operations, the more consequential question is not whether trade tensions will break the system, but where the cost of managing that uncertainty ultimately lands.

Consumers absorb what supply chains smooth over

Keeping the shelves stocked is a positive outcome, but it comes at a price. The ingenuity required to keep goods flowing in a volatile trade environment incurs costs, some of which have to be passed to consumers.

We recently surveyed 14,000 shoppers across seven major economies and found that nearly three in four (74%) had been affected by rising prices, with more than half switching to lower-cost brands as a result. Much of that pressure builds at the handoffs between suppliers of materials, components and intermediate goods, leading to incremental cost increases at each step.

From a consumer perspective, the distinction between policy decisions and business actions is blurred. When prices increase, shoppers assign blame broadly, to governments and brands alike. Trust is already fragile, so rising prices tied to trade uncertainty create reputational risk for businesses as well as political risk for governments. For supply chain leaders, this introduces a challenge that goes beyond resilience and efficiency.

Managing friction in the system

The way supply chain leaders and policymakers respond to uncertainty is diverging. Operations teams are planning for multiple scenarios, avoiding overreaction, and staying focused on maintaining the flow of goods. Political discourse on the other hand is leaning toward positions that challenge stability.

None of this is to say that trade policy does not matter, or that delays and constant changes are harmless. Over time, uncertainty acts like friction in the system. It slows investment and raises costs. The distinction is that the damage is subtle and slow-moving.

As geopolitical turbulence continues in 2026, most supply chains will continue to function. The challenge for businesses will be managing the downstream consequences of higher costs, and communicating effectively about them. This is not an easy conversation. Many consumers are operating with little tolerance for further price increases and are quick to assign blame when they occur.

The test for leaders is whether they can apply the same foresight to managing costs and explaining them, as they have to keeping supply intact and shelves stocked. Justifying the cost of resilience may soon matter as much as the operational excellence that delivers it.

About the author

Kevin O’Marah is the co-founder of Zero100, a London-based research services firm. Prior to helping launch Zero100, O’Marah was a director at Amazon, chief content officer at SCM World and a group vice president at Gartner. He was a Senior Research Fellow in Stanford University’s Graduate School of Business.

For more on this topic, read: Sustainability and AI: A complicated and often overlooked relationship