How is artificial intelligence being used by the world’s largest retailer? SupplyChainBrain recently engaged in a Q&A with Indira Uppuluri, SVP of Supply Chain Technology at Walmart.

Q: What prompted Walmart to explore the possibilities of AI in its supply chain operations?

Uppuluri: We see AI as a way to strengthen supply chain operations where precision, speed, and resilience are essential. By predicting demand more accurately, improving inventory decisions, and adapting quickly to disruptions, AI helps deliver a more seamless and dependable shopping experience for our customers. Equally important, we are people-led and tech-powered — our associates are always at the center of what we do. AI equips them with smarter tools that make their daily work more efficient and productive.

Q: What technologies did you apply, where, and how? What were the greatest challenges?

Uppuluri: Walmart operates one of the world’s most complex supply chains, serving customers who shop both online and in-store. With billions of possible delivery paths, we’re applying AI — from convolutional neural networks to reinforcement learning — to automate decisions end-to-end, turning scale, speed, and resilience into our greatest strength. 

We’re building several engines, powered by generative AI and machine learning, to operate behind the scenes to forecast demand, plan inventory placement, align labor and transportation, and keep products moving quickly through the network.

For example, our AI models can dynamically adjust how inventory flows. If an item starts selling faster in one region than another, supply can be shifted to meet that demand. When we ship from stores, our models dynamically decide which store in the customer catchment area can provide the fastest service with the fewest trips, ensuring the best possible customer experience. 

In our regional distribution centers, AI-powered robotics now break down and rebuild pallets, improving fulfillment accuracy and speed. On the logistics side, AI analyzes demand patterns, inventory locations, and real-time conditions to optimize delivery schedules and routes and to evaluate trade-offs in real time; routing algorithms maximize truck fill rates and select the most efficient pickup and drop-off points, reducing miles driven.

Q: Give me a sense of the "before" picture and the "after" picture.

Uppuluri: Typically, a supply chain runs on a 52-week prediction plan — built on the previous year’s purchasing patterns and adjusted for factors like weather and demographics. However, unforeseen events can upend that model overnight — COVID-19 is the clearest example. When customer behavior shifts in a matter of hours, a static outlook is no longer viable. Today, we also account for signals such as supplier manufacturing delays, route disruptions, site closures, labor and driver capacity, and shifting demand across geographies. To stay flexible, we need technology that can detect changes as they happen — and even anticipate them — so we’re ready for what comes next.

With generative AI and machine learning, that’s exactly what we do. We draw on a much broader set of real-time signals, adjust far more quickly, and use a digital twin of our supply chain to simulate “what-if” scenarios and stress-test decisions. Dynamic generative AI techniques such as retrieval-augmented generation (RAG) and multilayer perceptrons (MLPs) help our systems continually learn and surface opportunities we might otherwise miss. The result is a more agile, resilient network; shelves that stay stocked; and day-to-day operations that run more smoothly for our associates.

Read More: Walmart Touts Growth of AI Across Global Supply Chain

Q: In terms of learning curve, utilizing time, money and manpower, what applications of AI delivered the best bang for the buck so far?

Uppuluri: What’s important to us is not singling out one application but looking at how AI is helping us build a stronger, more resilient supply chain overall. We’ve invested in areas that make the greatest difference for our customers and associates — improving accuracy, streamlining processes, and giving our associates better tools to help them work more efficiently and more safely. At Walmart’s scale, even small improvements in efficiency or precision can translate into meaningful benefits.

Q: What plans do you have for further AI deployment in the future?

Uppuluri: Our journey is ongoing. We’re continuing to invest in AI, automation, robotics, and drones to shape a new era of commerce — one that gives customers the flexibility to shop however, wherever, and whenever they choose. The focus is on creating seamless, convenient, and personalized experiences for our customers while ensuring our associates are empowered with tools that make their work more productive, safer and more efficient. 

Q: What advice would you have for a company considering making the most of AI in its supply chain operations?

Uppuluri: Start with problems that truly matter at scale — AI delivers the most value when applied to well-defined challenges built on strong data foundations.

Bring your people along. At Walmart, we train associates to shift from being “players” who perform every task manually to “coaches” who guide and oversee with AI tools. That makes their work more efficient and frees them to focus on higher-value activities.

Think ahead. Customer expectations are rising, and they want things faster than ever. Our supply chain makes that possible, so investing in technology that drives speed and accuracy is critical.

Keep an open mind and focus on inventing the future!  What is standard operating procedure today, will likely be different in a year.