In The Arena by TechArena - AI Infra 2025: David Glick on Building Walmart’s Future with AI

Episode Date: November 2, 2025

In this episode of In the Arena, David Glick, SVP at Walmart, shares how one of the world’s largest enterprises is fostering rapid AI innovation and empowering engineers to transform retail.

Welcome to Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena. Welcome into the arena. My name's Allison Klein. We're coming to you from the AI Infra Summit in Santa Clara, California. And we have been talking to the most interesting people all week. I have a fantastic interview lined up up for everyone. We've got Dave Glick, Senior Vice President of Enterprise Business Services with us from Walmart. Walmart, while being known as one of the largest retailers in the world, is also a tech leader with a deep heritage in utilizing tech to transform business. Welcome to the program, Dave. Thank you, Allison. I'm glad to be here. So Dave, Walmart's never been on the show

before, but you guys don't really need an introduction. I think that one of the things that people don't know is the heritage of technology innovation at Walmart. I would love for you to share a little bit about IT practices at Walmart and what your role is as the Enterprise Business Services Senior Vice President at the company. If you think back all the way to like the 80s when Sam was leading the company personally, Walmart was a leader in technology. You know, one of the most famous things that they did is they drove adoption or we drove adoption of EDI over the years to make our supply chain more efficient

and working with suppliers. And that's just one example of the men. technology places where Walmart has led. Suresh Kumar joined us in 2019 to really supercharge the technology organization. And so you can think of the big fang companies as big tech, but we're right there. And we're not quite as big as them, but we've got thousands of engineers building stuff in-house. We work with vendors to bring things in-house. But we are applying technology to every aspect of what we do at Walmart.

My role is leading enterprise business services. So what is enterprise business services? on finance, tech and operations, people tech and operations, global governance, IT payments, and so on. I like to say that we pay our associates, we make the registers ring, and we close the books, everything you need to run the company. Just a few things under your purview. And I think that when we come to this AI moment, we're at AI InfraSummit, all of the things that you just described could be transformed and driven to higher productivity with AI. How are you looking at embracing AI in terms of driving real business value for Walmart? Yeah. So one of the great things about EBS or enterprise business services is we own the technology and the operations, for example, finance shared services or people services.

And that means that we can have the engineer sitting very closely with the operators and help them for their jobs. And my mantra is eliminate, automate, optimize. Wow. So we have processes all over the company. which we have humans doing, can we eliminate those processes upstream? If we can't do that, can we automate them? And if we can't do that, how can make the process for the humans most effective? We've been doing this for several years, but putting this on AI, bringing AI along,

puts this all on steroids. Now, when you think about AI in general, obviously traditional forms of ML have been around for a long time, generative AI, obviously a little bit less advanced, and then agenic AI coming on the scene. state of Walmart's total AI adoption. And how are you seeing that move forward in 2025? Yeah. You know, my boss, Suresh, likes to talk about building neural nets when he was in grad school in the last century. And that was the beginning of AI. But we've been doing AI for many years, both Suresh and I separately, but also at the company. And so it's transformed from AI to Gen.

which was a lot of chatbots. But I think the killer app is Agentic AI. Sure. And we've seen since agents launched and since the GEN-TIC AI launched, we have really dove deep. And so Doug has been saying, I joined the company two and a half years ago, and Doug's been saying since then, we need to do Gen A.I. We have questions about hallucinations and biases and how do we use it. And finally, earlier this year, we said, let's start building agents. That's awesome. I'm really glad that you brought this up because you were recently featured in the Wall Street Journal. You were talking a little bit about how you're deploying a GEN-TIC AI.

Tell me a little bit about the strategic driver for a unified AI framework and how super agents comes into the picture, because I love that concept. Yeah, we have a whole new vernacular. We have super agents and nano agents and chunky agents. But the idea is that we believe that we're going to have hundreds and thousands and tens of thousands of agents running around. Wow. And they will all do great things and humans will interact with all of those. We need a way to organize it. And so these super agents are three things, a single point of entry where you can ask it, you know, if you're in finance, you can say, let's see the sales yesterday. Or if you need to set up your direct deposit, there's an agent for that.

But you don't need to know the name or the URL for the direct deposit agent. You just go to the associate agent and it'll take you there. We've developed four super agents. One is Sparky. And Sparky is for consumers. The second is Marty, which is for suppliers, vendors, sellers, advertisers. The third is Wi-B, or Vi-Bi, which is like vibe coding for our developers. The fourth is my associate agent, which I think is the most diverse and the most interesting,

because it allows you to do your job, plus also handle all of the work-life commitments you have that you do on the network. I'm glad you brought that up. Now, let's click down into the Associate AI. When you look at how that interfaces with your Walmart employee workforce, how are you seeing productivity go up? What has the response been? Tell me a little bit about the experience with us.

Yeah, we are right now in a very experimental stage, where six months may seem like a long time was when we started getting to the point of like, let's just build stuff and see what happens. That's like 10 years now. Yeah, and so I'm like six months and we haven't seen productivity go up. But I think you're seeing it in little places.

And one of the things that was just an article from MIT saying 95% of AI POCs don't pay off, but a lot of those were sales and marketing. Sure. But actually the place where they say it is paying off is in back office functions, which is what we're doing. Makes perfect sense. And, you know, we have engineers who are building AI agents to automate things.

But the more interesting thing is we have actual non-engineers who are building things. Wow. And so I've been saying for a decade, we need to bring the engineers close to customers. But if the customer and the engineer are the same person, that's as close as you can get. Oh, wow. Yeah. And so we started with engineers building agents, which made their lives. better. And so now we're going to the non-engineers and saying, you can build agents to make your

lives better, too. And so that's where the magic is, which makes it even more important to have these super agents, because you're going to have this federated model of thousands, if not tens of thousands of people building agents, but you have to organize in some way, shape, or form. Now, Walmart, incredibly diverse functions in terms of your business, and then a vast consumer experience that you're driving, with people that come into your stores and visit you online. How do you see the human element of agenetic computing and driving broad proliferation of these agents in terms of adoption accelerating? And how are you thinking about tackling that as a team? Yeah, Walmart is a people-led tech-powered Ami Channel retailer. So people is at the part of what we do

every day. And one of the things we've found, especially in the back office systems, is that we've got people doing things which are wrote and repetitive and not super creative. And so they would much rather be spending their time either in the stores with associates or doing creative tasks. And so removing the toil is one of the things we say. Oh, wow. How do we remove toil from our engineers, from our associates and our customers and allow them to do more interesting things? I love that. Now, you've got some experience under your belt. You obviously have a ton of experience and working with engineering teams and knowing what's going to be a challenge to deploy. What do you think are the biggest pitfalls enterprises face as they start making this move from

thousands of POCs into broad deployment? Yeah, you know, I think everybody's a little skittish. They're like, how do we govern this? And, you know, we spent two years trying to figure out, like, how do we manage hallucinations, how do we manage biases, how do we have the right governance processes? And finally, we said, let's start building. We have lots of governance processes that we use for non-AI. We don't need special laws for AI. We have all kinds of laws already. But, like, let us use those processes. We have whole organizations that are in charge of keeping us safe. And so let us pressure test those.

Actually, the head of our audit team was like, hey, you know, I downloaded this thing and I started writing Python, but you guys wouldn't give me an API key. And we're like, that's on purpose. Because we have the right guardrails and the right mechanisms, which allow us to keep ourselves safe while going fast. That's a great way to think about it. Thank you so much for sharing that.

I think that when I think about where you're going, I can't wait to hear what happens in a year. You've been in six months. I can't wait to like come back to this event next year and say, how did the superagents go? Do you have more? All sorts of questions that I can think about asking you. But for right now, if you look forward three years, and I know that's hard. But if you look forward three years, what are you excited about most in terms of this AI journey at Walmart? Yeah, I know.

I try never to look forward three years. Maybe that's a defect on my part. But like, I look forward three months or even three weeks. Every Monday morning, we get a list of everything that launched over the last week. And it is a lot. There's new models. There's new startups. There's new stuff we're doing.

One of the things that I would like to leave the listeners is when we finally told the engineers, start doing stuff. Yeah. We don't even care exactly what it is, but start doing stuff and learning. And they came in and said, I build this thing in a week. And next Monday, you come in and say, oh, it was all terrible. I'm going to start over again. And I've learned so much in the last week. Now I can build it in three days, then I can build it in one day.

So the pace of progress, both in the LLMs and the supporting tooling, as well as our engineers, all of those are huge tailwinds. And so I would love for in three years us to think about all the toil we used to do and how it's gone. So we're allowed to innovate and be creative on behalf of associates and customers every day. Thank you so much for being on the show today. And thank you for being willing to share your vision in your journey because I think it's informing so many other organizations that are on the same journey path. For folks who want to engage with you, continue the story, ask questions, where can they go to learn more about what Walmart's doing and connect with you? Yeah, the best place to connect with me is on LinkedIn. I'm pretty active there.

You'll see me post a few times a day or excuse me a few times a week. Not a few times a day. You got to get an agent on them. We actually have Dave GPT, which helps with the rating. But that's the best place. And then obviously, we are posting all of the information about our agents on LinkedIn. And so that's the best place to find us. Awesome.

Well, thank you, Dave. Thank you so much for spending some time at AI Infro with us. Thanks for joining Tech Arena. Subscribe and engage at our website, Techorina.a. All content is copyright by Tech Arena. Thank you.