The conversation has moved beyond whether AI has potential. Most organizations already understand the opportunity. The real challenge is turning AI investments into measurable business outcomes. 

That’s where many companies get stuck. 

Not because they lack technology or ideas, but because the traditional delivery model can’t keep pace with the speed today’s business environment demands.  

Too many Salesforce initiatives take longer than expected, cost more than planned, and delay value realization. AI may have changed what’s possible, but many organizations are still implementing Salesforce the same way they did five years ago. 

At a time when executive teams are demanding faster ROI, that approach no longer works. 

A Different Era of Salesforce Delivery  

At Perficient, we believe the biggest opportunity in AI isn’t simply building AI-powered experiences. It’s transforming how those experiences get delivered. 

That’s why we’ve fundamentally reimagined our Salesforce delivery model around AI.  

Rather than adding AI to existing processes, we’ve embedded agentic intelligence throughout the software development lifecycle, accelerating discovery, design, development, testing, and deployment. 

The result isn’t incremental improvement but a fundamentally different way to delivery Salesforce transformation. 

Organizations experience the difference from Day 1: 

• Lower cost, with predictable engagement models focused on outcomes rather than hours. In one recent engagement, our approach helped a technology client avoid more than $1M in anticipated services costs. 
• Faster time to value, with delivery timelines accelerated by 30-50%. 
• Proven business impact, accelerating everything from proposal development to implementation and helping clients realize value sooner. 

This is why Perficient has become the Salesforce partner organizations turn to when speed, cost, or complexity are on the table. 

When executive expectations are high. When timelines are aggressive. When AI initiatives need to move beyond pilots and into production. When enterprise complexity threatens to slow progress. 

We help organizations move from vision to value without compromising the governance, scalability, or sophistication required by global enterprises.  

The Next Frontier: Accelerating Business Outcomes 

Technology alone doesn’t create competitive advantage. The ability to implement, scale, and operationalize it does. 

Because we’ve transformed how Salesforce gets delivered, we’re able to help clients tackle some of their most important business priorities faster, whether that’s deploying production-ready Agentforce experiences, modernizing revenue operations, or activating customer data to power more personalized engagement across channels. 

Our expertise spans the Salesforce platforms where organizations are making their biggest investments and expecting their biggest returns. 

Building the Agentic Front Office 

One example of that vision is what we call the Agentic Front Office.  

We believe organizations are entering a new era where employees spend less time navigating systems and more time directing outcomes. 

Instead of manually searching for information, triggering workflows, or coordinating across disconnected platforms, AI agents will orchestrate much of that work behind the scenes. Salesforce evolves from a system of record into an intelligence and orchestration layer that provides marketing, sales, and service interactions.  

The result is a front office that can move faster, respond smarter, and create more personalized experiences throughout the customer lifecycle. 

Headless 360: Reimagining the Experience Layer 

We’re also helping organizations rethink where those experiences live.  

As a Salesforce Headless 360 Launch Partner, Perficient helps clients design customer experiences around users, workflows, and AI agents rather than traditional application boundaries.  

The question is no longer, “What systems should users work in?” It’s become “What’s the fastest path to the right outcome?” 

Whether that experience happens inside Salesforce, a custom application, or a digital channel, Salesforce serves as the data, workflow, and orchestration engine that powers it all. 

See the Difference at Dreamforce 

This year, our Dreamforce story isn’t simply about AI.  

It’s about what happens when AI transforms how Salesforce gets delivered and helping organizations move from idea to implementation faster, reducing the time between investment and value, and proving there’s a better way to build, scale, and optimize Salesforce in the age of AI. 

If you’re looking for a partner that can help you accelerate outcomes, reduce complexity, and operationalize AI at enterprise scale, we’d love to show you what’s possible. Because the future isn’t just AI-powered experiences – the future is AI-powered delivery. And that future is already here.  

Connect with us at Dreamforce at one of our various events or speaking engagements: 

• Executive Strategy Sessions & Client Meetings | Available throughout Dreamforce  
• Welcome Reception | Monday, September 14 
• Agentforce Champions Breakfast Panel | Tuesday, September 15 
• TMT Happy Hour | Tuesday, September 15 
• Demo Open House | Wednesday, September 16 
• TMT Commercial Pre-Dreamfest Happy Hour | Wednesday, September 16
• Agentic Front Office From Pilot to Scale with Agentforce Session | Thursday, September 17 

The post The Future of Salesforce Isn’t Just AI. It’s How AI Changes Delivery. appeared first on Perficient Blogs.

Two Very Different Approaches to Enterprise RAG

In Part 3, we covered Azure — the natural choice for Microsoft-heavy enterprises. Today we explore two more platforms: AWS Bedrock and Oracle Cloud Infrastructure (OCI). Each has a distinct architectural philosophy and a clear ‘best fit’ use case.

AWS Bedrock: Maximum flexibility. The broadest model catalog in the industry. Built for teams that want to experiment with multiple LLMs and use managed services to minimize operational overhead.

OCI Vector Search: Maximum efficiency for Oracle customers. Native vector support directly in Oracle Database 23ai — no separate vector database needed. SQL-native vector queries from day one.

Choosing a cloud for RAG is not about which is ‘best’. It is about which fits your existing infrastructure, team skills, and cost model.

AWS Implementation — Amazon Bedrock + OpenSearch

AWS Services

AWS Architecture Flow

• Documents in S3 trigger a Lambda function via S3 Event Notification
• AWS Step Functions orchestrates the pipeline: Extract → Chunk → Embed → Store
• Chunking Lambda uses LangChain RecursiveCharacterTextSplitter (512 tokens, 50-token overlap)
• Amazon Bedrock Titan Embeddings V2 generates a 1024-dim vector per chunk
• OpenSearch k-NN index stores chunk text + vector + metadata using HNSW algorithm
• RAG app (Lambda + API Gateway): embed query → OpenSearch k-NN → build prompt → call Bedrock Claude → return answer
• CloudWatch captures all logs, latency metrics, and fires alerts on anomalies

AWS Bedrock — The Multi-Model Advantage

The biggest differentiator of AWS Bedrock is its model catalog. No other platform offers this breadth through a single unified API:

Bedrock lets you A/B test different LLMs with zero infrastructure changes. Switch Claude to Llama by changing one line — the API contract stays identical.

Bedrock Knowledge Bases — Managed RAG

For teams that want to move fast, AWS Bedrock Knowledge Bases is the fastest path to production RAG:

• Point it at an S3 bucket — AWS handles chunking, embedding, and OpenSearch automatically
• No infrastructure to manage — fully serverless, auto-scaling
• Integrated with all Bedrock LLMs — one API call returns a grounded answer with citations
• Trade-off: less control over chunking strategy, embedding model, and retrieval tuning

Use the managed option for prototypes and internal tools. Use the custom Lambda + OpenSearch path for production systems where retrieval quality tuning is critical.

AWS CodePipeline Testing

OCI Implementation — Oracle Database 23ai as Your Vector DB

Oracle Cloud Infrastructure takes a fundamentally different approach. Instead of adding a new vector database service alongside your existing infrastructure, OCI embeds vector capabilities directly into Oracle Database 23ai.

If your enterprise already runs Oracle ERP, Oracle databases, or Oracle Fusion middleware — you already have a production-grade vector database. You just need to enable it.

OCI Services

OCI Architecture Flow

• Documents uploaded to OCI Object Storage trigger an OCI Function via OCI Events Service
• Function extracts text and sends to chunking logic (Python-based splitter)
• OCI Generative AI Embed Models generate 1024-dim vector per chunk
• Chunk text + vector stored directly in Oracle DB 23ai using the native VECTOR column type
• RAG app: embed query → SQL VECTOR_DISTANCE() search → build prompt → call OCI GenAI → return answer

OCI’s Killer Feature — SQL-Native Vector Search

This is what makes OCI genuinely unique. Vector search in Oracle DB 23ai is just a SQL query:

SELECT chunk_text,

       VECTOR_DISTANCE(embedding, :query_vector, COSINE) AS similarity_score

FROM   documents_chunks

WHERE  category = 'HR'

ORDER  BY similarity_score

FETCH  FIRST 5 ROWS ONLY;

Notice the WHERE clause — you can combine traditional SQL filters with vector similarity in a single query. This enables hybrid filtering that other vector databases require complex workarounds to achieve.

AWS vs OCI — Side-by-Side Comparison

  In Part 5 — the final post — we cover the full testing strategy for RAG systems: the 4-layer testing pyramid, RAGAS metrics, hallucination detection, and the complete GenAI CI/CD pipeline.

The post RAG on AWS Bedrock vs Oracle Cloud: Architecture, Trade-offs, and When to Choose Each appeared first on Perficient Blogs.

AI is moving into the systems where banking work happens. That shift creates a new operating requirement for financial institutions: define what an AI agent can access, what it can do, and how every action will be governed. 

Model Context Protocol, or MCP, is part of that shift. Anthropic introduced MCP in November 2024 as an open standard for connecting AI assistants to systems where data lives, including business tools, development environments, and content repositories. Its purpose is to replace fragmented custom integrations with a consistent protocol for connecting AI systems to data sources and tools. 

For financial services, governed AI in banking cannot stop at drafts, summaries, and recommendations. AI needs a secure way to retrieve customer context, interact with workflows, and operate inside systems of record. MCP can serve as a standardized execution layer that allows AI agents to retrieve context, invoke actions, and execute workflows within established policies and audit trails.

Before execution scales, banks need to define control.

Governed AI in Banking Starts with Execution Boundaries 

A banking agent may need account context for a service interaction. That does not mean it should update customer data, initiate a loan workflow, trigger a fraud review, or close a complaint. Each action has a different risk profile. 

Before an AI agent acts inside a system of record, a bank should define six boundaries: 

• Data access: What customer, account, transaction, case, or relationship data can the agent retrieve? 

• Workflow rights: What approved processes can the agent invoke? 

• Human approval: What actions require review before execution? 

• Decision limits: What tasks remain advisory because they involve credit, compliance, fraud, or customer harm risk? 

• Exception handling: What happens when data is incomplete, conflicting, stale, or high-risk? 

• Evidence capture: What gets logged for audit, compliance, model governance, and operational review?

These boundaries define the operating model for governed AI in banking. MCP can standardize how agents connect to enterprise capabilities, but the institution must decide what those agents are authorized to do. 

Where Execution Boundaries Become Enforceable 

Execution boundaries only matter if they can be enforced inside daily workflows. For banks using Salesforce, the platform can serve as part of the control plane for agentic banking work, connecting identity, permissions, customer context, workflow logic, action history, and human review.

Salesforce can serve as an API-first foundation for modern banking delivery, providing a secure, auditable execution layer built for governance, guardrails, and human oversight.

MCP extends that foundation by standardizing how agents connect to approved tools and resources. With Agentforce, organizations can register and manage MCP servers so agents can securely access approved tools and data. For banks, the server boundary is the control point: the agent should only see the capabilities its role is permitted to use. 

Governed MCP Requires Production-Level Discipline 

Once execution boundaries are defined, banks need to translate them into controls that work inside daily operations. 

• Make agent identity explicit. Define whether the agent acts for a service representative, advisor, operations team, compliance function, or automated process. Do not let it operate as a generic system user. 

• Scope permissions to the workflow. A servicing agent may retrieve account context. A lending agent may prepare a document checklist. A compliance agent may flag missing evidence. Each agent should access only what its role requires. 

• Set enforceable decision thresholds. Credit decisioning, fraud escalation, KYC and AML screening, complaint handling, and exception processing need clear stop points. Some actions can run automatically. Others should route to a person. 

• Capture evidence during execution. MCP-connected AI can generate change logs or audit trail entries as part of the action, reducing the need to reconstruct documentation later. 

• Monitor after launch. Banks need visibility into agent performance, failures, overrides, and exception patterns so governance improves with real operating data. 

Regulatory Expectations Make Evidence Central 

Financial institutions remain accountable for AI-enabled decisions and workflows. The Consumer Financial Protection Bureau (CFPB) has stated that creditors using complex algorithms, including AI-driven credit models, must provide specific and accurate reasons for adverse action. Vague explanations that obscure the actual basis for a decision do not meet that standard.  

In April 2026, the Federal Reserve, Office of the Comptroller of the Currency (OCC), and Federal Deposit Insurance Corporation (FDIC) issued revised model risk management guidance that replaced SR 11-7 and emphasized risk-based governance, validation, monitoring, controls, and third-party considerations. The guidance notes that generative and agentic AI are still evolving, and it is expected to be most relevant to banking organizations with more than $30 billion in total assets. 

For banks, the regulatory signal points back to the same operating requirement:  

• Agentic AI in banking must be explainable, constrained, monitored, and auditable. 

• MCP creates a cleaner connection layer between AI and enterprise systems. 

• Salesforce provides the governed foundation for access, workflow, identity, and action history. 

• AI-First Engineering brings the delivery discipline to connect architecture, integration, testing, security, compliance, and adoption. 

The institutions that get value from MCP will define safe execution before they scale governed AI in banking.

The post Governed AI in Banking Starts with Defining What Agents Can Do appeared first on Perficient Blogs.

AI productivity claims usually focus on speed: faster documentation, summaries, planning, and releases. In banking, speed only creates value when the work still meets the standards of risk, compliance, security, operations, and audit. 

The real friction often starts as work moves from one team to the next: 

• A developer closes a story, but release notes still need to be assembled from tickets, commits, and change records. 

• QA completes testing, but evidence still needs to be formatted for compliance review. 

• A product owner prepares for sprint planning, but prior sprint data must be rebuilt from status updates and carryover work. 

• A release is technically ready, but risk, security, operations, and business teams still need a complete record of what changed. 

Those handoffs create room for error. A release note may not match the actual change. A compliance reviewer may ask for missing test results. A deployment may wait while teams track down an approval, dependency, or change record. What begins as missing context becomes rework. 

This is not administrative waste. In regulated delivery, coordination is part of the control environment. The problem is how much manual effort it takes to keep that environment moving. Many financial institutions face the same pattern: duplicated logic across channels, manual status tracking, and developers spending time integrating systems instead of building value.

How MCP Productivity in Banking Reduces Rework 

Model Context Protocol, or MCP, gives AI assistants a standard way to connect to enterprise systems and data sources. For banking delivery teams, the value is practical: AI can work from live, permissioned context instead of disconnected prompts, stale exports, or manually assembled status reports. 

That is where MCP productivity in banking starts to create value: by reducing delivery friction, giving AI agents standardized access to enterprise context, and invoking approved capabilities through governed systems. In a Salesforce environment, that context can include delivery activity, customer data, workflow history, approvals, and related records. 

Salesforce can serve as the system of action where delivery context, customer data, and workflow activity and approval history are governed. Agentforce extends that foundation with an MCP client, allowing organizations to register and manage MCP servers so agents can access approved tools and data through a managed experience.  

The result is a cleaner evidence flow. Delivery artifacts can be generated from governed workflow data instead of rebuilt from scattered updates after the fact. An MCP-connected agent can pull from approved sources, apply a consistent structure, flag missing evidence, and route the output for review. 

A human still owns the final artifact. The difference is that teams spend less time searching, reconciling, and rebuilding context before the next handoff can happen. 

How to Measure MCP Productivity in Banking 

Organizations implementing MCP-enabled workflows, often see measurable reductions in manual effort in areas such as:

• Release notes and sprint summaries: Drafted in minutes instead of one to three days of manual aggregation. 

• QA and compliance documentation: Completed 25 to 44 percent faster, with more consistent artifacts. 

• Sprint planning preparation: Reduced from 30 to 45 minutes of manual assembly to minutes using prior sprint data. 

• Delivery flow: 20 to 35 percent improvement in cycle time, throughput, and predictability. 

The delivery flow improvement is the number to study. It shows whether work is moving through the system with fewer delays, fewer missing inputs, and fewer avoidable review cycles. 

Leaders should evaluate MCP productivity in banking by mapping where work stalls today: 

• Where do teams manually recreate status? 

• Where is compliance evidence missing or inconsistent? 

• Which release steps depend on one person assembling information? 

• Which workflows require the same data to be collected more than once? 

A use case that only makes one person faster may stay local. A use case that reduces review cycles, rework, evidence gaps, or queue time can improve the delivery system. 

That is the larger opportunity behind MCP productivity in banking. Salesforce provides the workflow foundation. MCP connects agents to approved systems and context. Perficient’s AI-First Engineering brings the discipline to integrate governance, identity, testing, compliance, and adoption into production. The result is less rework, less reconstruction, and more predictable flow. 

The post MCP Productivity in Banking Comes from Less Rework appeared first on Perficient Blogs.

In a recent blog, we explored how Brazil’s domestic instant payment platform, Pix, is reshaping digital commerce. This article builds on that discussion by examining the next opportunity for payment leaders: extending the convenience of domestic payment platforms across borders.

In speaking with payment rails executives globally, we were asked to examine how a domestic-only payment platform can expand to offer outbound cross-border QR-payments. For payment leaders, this expansion requires more than enabling international payments. It requires the right partnerships to extend global reach maintaining security, scalability, compliance, and a familiar customer experience.

How Modern Payment Platforms Scale Internationally

Many payment rails executives will immediately think of Wero, which, for readers who are unfamiliar, is a pan-European mobile payment system launched by the European Payments Initiative (EPI). Wero allows users to send and receive money instantly using a phone number or QR code, without entering the recipient’s International Bank Account Number, or IBAN.

Money moves directly from bank account to bank account via Single Euro Payments Area (SEPA) instant payments. However, Wero did not start as a domestic-only platform that later expanded across borders. According to the EPI, Wero was designed as a pan-European payment solution from the outset, with a phased rollout beginning in Belgium, France, and Germany in 2024.

Another example payment rails executives are encouraged to consider is barq, a Saudi financial application. Startup Researcher reports that barq reached 1 million users within 21 days of launch and surpassed 10 million users within 17 months. Barq recently expanded from domestic payments to outbound cross-border QR payments through an integration with Alipay+, Ant International’s global payment gateway.

According to the April 2026 announcement from Alipay+, the agreement makes barq the first Alipay+ payment partner in the Middle East to enable outbound cross-border QR payments for more than 12 million users from Saudi Arabia across more than 220 markets worldwide.

The integration shows the value of partnerships. Rather than building individual connections in each country, barq leverages the Alipay+ infrastructure to offer its customers widespread acceptance across international markets.

Supporting Secure Payment Growth

As payment platforms expand their services, they need technology that supports secure and reliable payment processing. In September 2024, Mastercard announced an agreement with barq to provide payment acceptance solutions through Mastercard Gateway. This technology supports recurring and scheduled payment needs, including subscriptions, instalment plans, bill payments, and automatic transactions that do not require the customer to be present on the website or app. These capabilities provide greater convenience, speed, and security for merchants and users.

Mastercard Gateway provides the payment processing and fraud prevention capabilities behind the barq’s expanding payment services. Behind these customer-facing capabilities, a broader cloud foundation supports growing transaction volumes, real-time performance, and continuous availability.

Scaling the Payment Platform Through Cloud

While Mastercard Gateway powers the customer’s payment experience, Google Cloud provides the underlying technology foundation that enables the platform to scale securely, process transactions reliably, and support continued innovation.

Google Cloud provides a scalable, secure foundation that supports millions of active wallet users while helping the platform meet evolving AI, compliance, and data security requirements.

Google Kubernetes Engine (GKE) supports real-time container orchestration. GKE automates, monitors, and scales financial microservices. Together, Google Cloud and GKE give technology teams the flexibility to deploy updates quickly, respond immediately to changing customer needs, and maintain performance as transaction volume grows.

From Local Platforms to Global Payment Ecosystems

Digital wallets are becoming part of a more connected global payments ecosystem. As acceptance networks expand across markets, payment providers have an opportunity to extend familiar payment experiences into new regions while giving merchants more ways to serve international customers.

Participating in this ecosystem requires secure payment capabilities, connections to global networks, and cloud infrastructure that supports real-time transactions at scale. Together, these elements allow payment providers to pursue international growth while maintaining the speed, reliability, and convenience users expect.

Conclusion

Consumers increasingly expect their preferred payment methods to work across markets, while merchants seek efficient ways to serve international customers. Meeting both expectations takes a connected ecosystem of secure payment capabilities, global acceptance networks, and scalable cloud infrastructure.

Strategic partnerships help digital wallet providers bring these components together and pursue international growth without building and managing separate connections in every market.

As a Google and Mastercard partner with payment technology and cloud expertise, we help providers build the connected ecosystems needed to scale securely across borders. Contact us to discuss your cross-border payment strategy.

The post The Next Phase of Payments: From Domestic Success to Global Scale appeared first on Perficient Blogs.

Data warehouse modernization has long been one of the most complex initiatives organizations undertake. While moving data itself is often straightforward, migrating decades of embedded business logic, stored procedures, ETL workflows, reporting dependencies, and platform-specific SQL introduces significant cost, risk, and complexity.

The post Accelerating Data Warehouse Modernization with Snowflake AIM appeared first on Perficient Blogs.

This achievement reinforces our commitment to helping clients harness the power of AI to drive innovation, improve operational efficiency, and create exceptional customer experiences. As organizations increase their AI investments, we bring together deep industry expertise, strategic consulting, and advanced cloud capabilities to deliver measurable business outcomes at scale.

Recognized for Proven AI Expertise

The AWS AI Competency validates Perficient’s ability to design, build, and deploy AI-powered solutions that address complex business challenges across industries. The designation highlights our experience helping clients leverage AWS AI services, including Amazon Bedrock, Amazon SageMaker, and Amazon Q, to:

• Accelerate generative AI adoption and innovation
• Modernize customer and employee experiences with intelligent solutions
• Automate business processes and workflows
• Enhance data-driven decision making
• Scale AI initiatives securely and responsibly

AWS AI Competency Partners undergo a rigorous validation process that evaluates technical proficiency, customer success, and adherence AWS best practices, including the AWS Well-Architected Framework. Achieving this distinction demonstrates our ability to deliver transformative AI solutions that help organizations realize value faster.

Strengthening Our Collaboration with AWS

Perficient and AWS share a commitment to helping organizations unlock meaningful business outcomes through emerging technologies. Achieving the AWS AI Competency further strengthens our collaboration and positions us to help clients adopt and scale AI with confidence.

“Perficient achieving the AI Competency is an incredible achievement,” said Steve Holstad, VP of AWS Practice at Perficient. “Perficient and AWS are both deeply invested in helping clients unlock business outcomes through AI, making this a perfect time to strengthen our collaboration. We greatly appreciate the partnership and are excited for what’s ahead.”

Delivering AI Innovation Across Industries

Organizations across every industry are exploring how AI can improve productivity, enhance customer engagement, and create competitive advantage. Perficient helps clients navigate this evolving landscape by combining AI strategy, cloud engineering, data modernization, and industry expertise to deliver practical solutions aligned to business objectives.

From generative AI applications and intelligent automation to advanced analytics and machine learning initiatives, we help clients move from experimentation to enterprise-scale adoption.

This expertise is already delivering measurable results for clients. For example, Perficient helped a leading U.S. life insurer modernize its customer service knowledge platform with an AWS-powered generative AI solution. The platform scaled to more than 1,100 active users, answered over 110,000 insurance policy questions in its first three months, and helped reduce customer service call handling times by 13%.

In another engagement, Perficient helped a global manufacturing leader deploy generative AI-powered customer service capabilities on AWS that deflected 975,000 customer interactions in the first year, generating approximately $25 million in annual cost savings while improving support experiences at scale.

Looking Ahead

As an AWS AI Competency Partner, Perficient is well positioned to help clients accelerate their AI journeys and transform ambitious ideas into measurable business outcomes.

We are honored to receive this recognition and look forward to continuing our work with AWS to help organizations innovate responsibly, scale confidently, and capture the full potential of AI.

Learn more about Perficient’s AWS capabilities and how we help organizations drive business transformation through cloud and AI innovation.

The post Perficient Achieves AWS AI Competency, Accelerating AI-Powered Business Transformation appeared first on Perficient Blogs.

Our first profile features Tauni Crefeld, Head of Customer Success, who makes an impact through leadership, problem-solving, and a deep commitment to helping people succeed. Whether she’s coaching teams through complex delivery challenges or sparking conversations through storytelling, Tauni brings a thoughtful, people-first approach to everything she does.

In our People of Perficient profile, Tauni shares her approach to leadership, what drives her work, and how her experiences — from the Air Force to consulting — continue to shape her perspective today.

What is your role at Perficient?  

At Perficient, my technical title is Head of Customer Success. I focus on helping ensure our complex and fixed-fee projects are successful.

I think about the role in three areas. The first is making sure projects are set up for success by reviewing estimates, solutions, and contracts — I read every fixed-fee percent complete contract before approval to make sure it’s set up correctly. Second is supporting and coaching teams during delivery of fixed-fee and complex projects. The third is helping teams successfully close projects while maintaining strong client relationships.

How have you grown as a leader throughout your career?

My first job was in the Air Force. I was Air Force Security Police at three different military bases. I came in as a lieutenant and oversaw 40 people, all men, almost all older than me.

So, when you come in as a 21-year-old female, and you’re in charge but don’t know anything yet, you have to listen and find a way to lead. You learn quickly how to get along, build relationships, and lead by influence because despite what you see in movies, you can’t just come in and start giving orders. It will not work.

When you were walking into that room as a 21-year-old, what was your headspace? 

It took about a week to even understand things.

I went on a training run with a senior lieutenant. We were out in the missile fields of North Dakota, in the middle of nowhere, going from launch facility to launch facility. He would make radio calls back to base control and show me exactly what to say. Then, he’d have me try it.

When I did, there would just be silence — no response.

He would take the mic back, say the exact same thing, and they would respond to him. After the tour of duty, we went back and asked base command why they hadn’t responded to me. And they said, “There are no female lieutenants here, so we thought it was a prank.”

There were things like that from the beginning. Walking into a room of 40 people is already intimidating, but walking in with that context makes it a little harder. I kept a veneer up for a couple of days trying to maintain this idea of what a “professional leader” should look like. And then I just dropped it. I knew it wouldn’t work.

I started asking questions, listening, and learning: “Tell me what you do. Tell me your role. How does this work?” From there, they adapted to me, and I adapted to them. It was an evolution, but I had to let go of that façade of what I thought leadership should look like.

How did your approach to leadership translate from the military into consulting?

After my military career, I spent almost 25 years at another consulting company. I started in an entry-level analyst role and worked my way up to managing director by taking on bigger and bigger projects.

I focused more on delivery than sales and made myself valuable through delivery and fixing messy teams. I didn’t go around doing traditional “selling,” but I sold through strong delivery. And I always say that good delivery can be a great sales tool.

At Perficient, I have a title, but I don’t use it much. Instead, I approach teams by saying, “I’m here to help.” I don’t rely on positional authority. People are open to help and want their projects to succeed.

What motivates you throughout the day?

Two things motivate me. I love coaching people. The people I’ve met here are incredibly smart and receptive to feedback. They’re eager to learn, and I love helping them grow.

The other thing I love is fixing problems. I even have a book called The Fixer. I like going into messy situations, figuring out what’s wrong, and changing the dynamics so teams can deliver well. My goal is always to make things better for everybody involved.

What do you enjoy the most about working with different teams and cultures?

I love working with people around the world. I speak Spanish, and I love connecting with our colleagues in Latin America and using the language whenever I can.

The best thing about meeting so many different people is that everyone brings a different perspective to what we’re doing. When I meet a new team, I don’t ever include my title. I ask questions, listen, and try to understand their situation first. I like learning and listening. You can’t fix a problem if you don’t understand what’s going on.

In your time at Perficient, how have you engaged with the Women in Technology ERG?

I was invited to speak about my book, Own Your Path, which focuses on driving your career forward. The Women in Technology ERG has also hosted a book club around it, and I’ve attended several sessions with them.

I had all of these stories in my head, so I started writing them down. I figured that a lot of the young women that I’d worked with back at my other company would be the ones who read it.

The whole goal of these stories is to help people, so the stories are very solution-oriented — the same way I try to fix projects: Find all the commonalities and structural problems and then find ways to overcome them. So, it’s amazing to see people read it and take some of those learnings to heart. That makes me so happy because that’s the whole point of the book.

How can we elevate women working in technology?

One thing I appreciate about WiT is that it’s not exclusive to women. I have seen men participate in those meetings, and that’s important. The only way women can be elevated is if we get everyone in the conversation.

I made it to the level that I did because I had good sponsors and supporters, all of whom happened to be male. I did have good women mentors, but it was the men who were able to push me forward because they have more positional authority. Right now, in the world we live in, it’s the men who are still in those top positions.

We need to get the men in the conversation. Women can’t solve this on their own because a lot of the challenges are structural.

What advice do you have for women working in tech?

Women listen a lot, and we’re very good at it. But many of us are not good at advocating for ourselves.

If there’s a skill that you want to learn or a position you want to grow into, then tell people. Tell people what you want. Otherwise, they may guide you in a different direction.

You have to decide what’s right for you and push ahead. They don’t know what you want. Only you do.

SEE MORE PEOPLE OF PERFICIENT

It’s no secret that our success is because of our people. Across teams, technologies, and time zones, our colleagues collaborate to build AI-first solutions that hold up in the real world and deliver outcomes that matter. We’re always looking for unfiltered, unconventional, and unshakable problem-solvers who turn bold thinking into results. Join us and be part of a culture defined by curiosity, ownership, and impact.

Visit our Careers page to explore open roles and learn more about what it’s like to work at Perficient. Join our talent community for career tips, job openings, company updates, and more!

Go inside Life at Perficient and connect with us on LinkedIn, YouTube, X, Facebook, and Instagram.

The post From the Air Force to Customer Success, Tauni Crefeld Champions Personal Growth appeared first on Perficient Blogs.

Why Azure is the Natural Home for Enterprise RAG 

If your organization runs on Microsoft — SharePoint, Teams, Azure AD, Microsoft 365 — then Azure is your natural RAG platform. Every data source your employees use daily connects natively, without custom connectors or complex authentication flows. 

Azure also provides something no other cloud currently matches: Azure AI Search with hybrid search — combining traditional BM25 keyword matching, vector similarity, and a semantic ranker in a single query. For enterprise content (policies, documentation, technical specs), this hybrid approach consistently outperforms pure vector search.

Azure AI Search’s semantic ranker uses a cross-encoder model to re-rank results after initial retrieval — dramatically improving precision for complex, multi-concept queries.

Azure Services at a Glance 

The Azure RAG Architecture — Step by Step 

Data Ingestion Flow 

• Documents uploaded to Azure Blob Storage trigger an Azure Function automatically 
• The Function extracts text using PyMuPDF (PDFs), python-docx (Word), or Graph API (SharePoint) 
• Text is cleaned, normalized, and split into 512-token chunks with 50-token overlap 
• Each chunk is sent to Azure OpenAI text-embedding-3-small — returns a 1536-dim vector 
• Chunk text + vector + metadata (source, page, date) is upserted into Azure AI Search index 

Query Flow (Every User Question) 

• User submits a question through the RAG application (web app, Teams bot, or API) 
• Question is embedded using the same text-embedding-3-small model 
• Azure AI Search runs a hybrid query: BM25 keyword match + vector similarity + semantic reranker 
• Top 5 most relevant chunks are returned with their source metadata 
• Chunks are injected into a structured prompt and sent to Azure OpenAI GPT-4o 
• GPT-4o generates a grounded answer — constrained to the retrieved context only 
• Answer is returned to the user with source citations 

Always instruct LLM to answer ONLY from the provided context and to say ‘I don’t have that information’ when the context doesn’t contain the answer.  This keeps responses grounded in verified source material and reduces unsupported AI-generated content.

Azure AI Search — Why Hybrid Mode Is Critical 

Pure vector search is powerful, but it has a known weakness: exact keyword matching. If a user types an exact product code like ‘PRD-2024-XA7’, vector search may not surface the right document because the embedding captures semantic meaning — not exact strings. Azure AI Search solves this with hybrid mode:

For enterprise use cases — especially HR policies, compliance documents, and technical specs — always enable hybrid mode with the semantic ranker. 

Azure DevOps CI/CD Pipeline for RAG 

A production RAG system needs automated testing on every code change. Here is the complete Azure DevOps pipeline: 

RAGAS Thresholds for the Azure Pipeline 

If any RAGAS metric falls below threshold, the pipeline fails and the deployment is blocked. This is your automated quality gate for every release.

Security Considerations on Azure 

Azure provides a mature security stack for RAG systems: 

• Azure Key Vault — store all API keys, connection strings, secrets — never in code or config files 
• Azure AD (Entra ID) — use Managed Identity for Function-to-Service authentication — no credentials needed 
• Azure AI Search — role-based access control via Azure RBAC — restrict who can query which indexes 
• Private Endpoints — keep all service communication inside your VNet — no public internet exposure 
• Azure Defender for AI — detect anomalous query patterns, prompt injection attempts 

Managed Identity over API Keys: Configure Azure Functions to authenticate to Azure OpenAI and Azure AI Search via Managed Identity. This eliminates credential rotation risk entirely. 

  In Part 4, we compare building RAG on AWS Bedrock vs Oracle Cloud Infrastructure — two very different approaches with distinct trade-offs for enterprise teams. 

The post Building Enterprise RAG on Azure: GPT-4o + Azure AI Search + Azure DevOps — End to End appeared first on Perficient Blogs.

RAG (Retrieval-Augmented Generation) is a technique that combines information retrieval with a Large Language Model (LLM) instead of asking an LLM to answer a question only from pretrained knowledge, RAG first retrieves relevant information from specific knowledge sources, the retrieved information is then provided to the LLM as context, which helps the LLM generate a more relevant and grounded answer.

Typical RAG flow:

User Question → Retrieve Relevant Chunks → Provide Context to LLM → Generate Answer

Why Do We Need RAG Evaluation?

Building RAG pipeline does not automatically guarantee that it will produce correct answers. There can be problems at different stages of the pipeline.

For example:

Problem 1 – Poor Retrieval

The question may be:

“What is the work-from-home policy?”

But the retriever may return chunks about leave policy instead.

The LLM now has incorrect or irrelevant context.

Problem 2 – Missing Information

The correct document may contain the answer, but the retrieval process may fail to retrieve the required chunk.

In this case, LLM may not have enough information to provide the correct answer.

Problem 3 – Unsupported Answer

The retrieved context may contain the correct information, but the LLM may generate an answer that includes information that isn’t present in the context.

This is commonly referred to as hallucination or lack of grounding.

Problem 4 – Irrelevant Answer

The retrieved context may be correct, and the answer may contain information, but it may not actually answer the user’s question.

Therefore, simply checking the final response manually isn’t enough.

We need a systematic way to evaluate different aspects of the RAG pipeline.

What is Ragas?

Ragas is an open-source framework for evaluating LLM applications, with strong support for Retrieval-Augmented Generation (RAG) pipelines.

Ragas provides evaluation metrics that help measure different aspects of a RAG system.

Instead of simply asking:

“Is this answer good?”

We can evaluate questions such as:

• Did we retrieve the right information?
• Did we retrieve enough relevant information?
• Is the generated answer supported by the retrieved context?
• Does the answer address the user’s question?

This allows us to identify where the RAG pipeline is performing well and where it needs improvement.

What We Are Going to Build

In this blog, we will use a PDF as the knowledge source. We will generate synthetic test data from the PDF using Ragas, run each generated question through the RAG pipeline, and use an LLM as a judge to evaluate the results.

We will evaluate four metrics: Answer Relevancy, Faithfulness, Context Precision, and Context Recall.

The overall flow is simple:

PDF →  Load PDF →  Chunk PDF →  Generate Test Data using Ragas  →  Run each question through RAG  →  LLM as Judge  →  Evaluation

1. Install the Required Packages

First, install the libraries required for this example.

pip install -U ragas langchain langchain-community langchain-openai langchain-text-splitters pypdf faiss-cpu python-dotenv

2. Load the PDF

Let’s assume our input document is called employee_handbook.pdf.
We first load the PDF using PyPDFLoader.

3. Chunk the PDF

A PDF can contain a lot of information. We split it into smaller chunks so that the RAG application can retrieve the relevant information when a question is asked.

4. Create the RAG Retriever

For our simple example, we create embeddings for the chunks and store them in FAISS (can use any vector DB here). The retriever will use this store to find relevant chunks for each question.

5. Generate Test Data Using Ragas

This is one of the useful parts of Ragas.

Instead of manually creating questions from the PDF, Ragas can generate synthetic test data based on the document content.

For example, Ragas can generate questions such as:
• What is the leave policy?
• How many vacation days are available?
• Who approves a leave request?

This gives us a test dataset that can be used to test our RAG pipeline.

6. Run Each Question Through the RAG Pipeline

Now we take each generated question and send it to our actual RAG application.

The RAG application retrieves relevant chunks and uses an LLM to generate the answer.

For each test question, we collect three important things:

• The question
• The generated answer
• The retrieved context

7. Use an LLM as a Judge

Now comes the evaluation part.

We use LLM as a judge. Its job is not to generate the RAG answer. Its job is to evaluate the answer produced by the RAG application.

The judge looks at the question, answer, retrieved context, and reference information as required by the metric.

8. Evaluate the RAG Using Four Metrics

To evaluate the RAG pipeline, we will use four metrics that assess the quality of retrieved context and generated answers.

    1. Answer Relevancy
What does it measure? Does the generated answer directly address the user’s question?

Example:
Question: How many vacation days are available?
Relevant answer: Employees receive 20 vacation days per year.

If the answer talks about unrelated employee benefits instead, the answer is less relevant.

In simple → Did the answer address the question?

    2. Faithfulness
What does it measure? Is the answer supported by the retrieved context?

If the retrieved context says employees receive 20 vacation days, but the RAG answer says 30 days, the answer is not supported by the context.

In simple → Did I answer using the retrieved information?

    3. Context Precision
What does it measure? Did we retrieve relevant information?

For a question about leave policy, we want the retriever to return leave-policy information rather than unrelated information such as parking or IT security.

In simple → Did I retrieve relevant information?

    4. Context Recall
What does it measure? Did we retrieve all the important information needed to answer the question?

If the answer needs information A, B, and C but the retriever only finds A and B, some important information is missing.

In simple → Did I retrieve all the important information?

9. Understand the Results

Suppose we get the following scores, we can understand them simply:

Answer Relevancy = 0.91 → The answers are mostly relevant to the questions.
Faithfulness = 1.0 → The answers are mostly supported by the retrieved context.
Context Precision = 1.0 → Most retrieved information is relevant.
Context Recall = 0.5 → Some important information may not have been retrieved.

The actual pass/fail threshold should be decided based on the requirements of your application.

Conclusion

Without an evaluation framework, we may test an RAG application by asking a few questions and manually checking the answers. Ragas provides a simple way to evaluate a RAG application using measurable metrics.

References

Ragas Documentation: https://docs.ragas.io/
Ragas Test Data Generation: https://docs.ragas.io/en/stable/concepts/test_data_generation/
Ragas Evaluation: https://docs.ragas.io/en/latest/references/evaluate/

The post Evaluating a RAG Pipeline Using Ragas appeared first on Perficient Blogs.