.png)
The Doing Business in Bentonville Podcast
To create an ecosystem that connects leaders of all kinds – industry, community, student, educational, civic, investment and entrepreneurial – to help overcome Omnichannel Retail barriers through exclusive, insight-rich content.
The Doing Business in Bentonville Podcast
Ep. 133 - Trust, Data, and the AI Shelf War
The retail playbook just changed: shoppers still click and scan, but AI agents now browse, compare, and buy on our behalf. We brought together leaders from academia, CPG, platforms, and agencies to break down what that means for brands selling at Walmart and across the modern digital shelf. The big takeaway is simple and hard: trust wins. Trust between people and machines, and trust between models and your product data. Clean attributes, consistent claims, and verifiable signals across PDPs, retail media, and third-party sources are now the difference between being recommended, or ignored.
We dive into the “perpetual moment of truth,” where inspiration, evaluation, and purchase collapse into one seamless flow. You’ll hear how bot-facing content, bullets, tables, certifications, and structured data, helps AI reason and cite your products. We explore generative engine optimization (GEO) versus traditional SEO, why Content Quality Scores keep shifting, and why Walmart’s new item setup will demand complete attribution at creation. Expect concrete tactics: set up agents to monitor rankings and data health daily; build a shared prompt library; reverse-engineer great answers; and run A/B tests that balance keyword search with conversational queries.
New modalities are already here: AI-enabled browsers, vision-plus-voice experiences, and wearable interfaces that parse shelves in real time. That makes consistency across packaging, imagery, and PDP claims critical, especially for safety and dietary needs. Our panel keeps it practical; scope one category, perfect attributes, verify claims, and let auditing agents alert you to drift. Move fast without breaking trust, and treat conversation as the new code your whole team can write.
If this helped you think clearer about AI in retail, subscribe, share with a teammate, and leave a review to help others find the show.
Good morning, everyone. Good morning and welcome to doing business in Bentonville AI Retail Summit. Thank you for being here. It's so good to see everyone. Uh, it's been almost five years since doing business in Bentonville has been able to have a live event. And I will tell you, I've been looking forward to this for the last five years. My name is Andy Wilson. I'm the executive director of doing business in Bentonville. And I'm going to tell you a bit about us and remind you about doing business in Bentonville. I gave you a little history, and then I want to tell you where we're going as a company. Um, I want to still I want to step back. I guess Cameron, our late Cameron Smith, our friend, founded doing business in Bentonville, I guess Wayne in about 1989, somewhere in that area, right? And so anyway, um six years ago, I get this call from uh Cameron Smith. And if you know the late Cameron Smith, it goes like this Andy, what are you doing? Well, I'm working. I want you to my office. Yes, sir. And that was it. I drove to his office. And now, if you know Cameron, you did not turn Cameron down, right? Because you went to his office or wherever when he called. So I sat down, I sat down with him, and then Wayne Callahan is here. Wayne, thank you, my friend, because we had a great conversation that day. And Wayne and Cameron said, Hey, Andy, we want you to lead doing business in Bentonville. And what we did back then, if you remember, if you were here then, we had live events here in Bentonville, Rogers at a hotel. We invite our supplier group, we had executives there from Walmart or from other organizations, and we talked about our business in retail. My my background is almost 30 years at Walmart. I grew up there. And so growing up there, I seen a lot of change, just like we have today. There's so much change in the in the future. We're going to talk a bit about that in a few moments. But it was so exciting. So we were doing the live events, then COVID happened. We all know what happened then. The live events stopped. During that time, I met a friend, Eric Hirrington. And Eric and I became very good friends, and he had a tech company at the time. And we just had conversations and coffees throughout. The only reason, though, that he spent time with me is because he wanted for the price of one, two tables, so he can invite people there. And I would charge him on for only one table. I've really I realized that at that point, but we still made continued our friendship. So anyway, our friendship grew. And then uh after Cameron passed away, Eric acquired doing business in Bentonville. And he called me up, and we I was working with him and Wade all through this all through the process. And uh uh Eric shared with me some ideas about the future. And he shared to me, he said, Andy, the future is podcast shows. He said, That's the future. And I knew nothing about podcast shows. All I would listen, I'll watch some podcasts from time to time, but I knew nothing about it. He said, I'm gonna teach you about the podcast shows. We're gonna develop a media company. Today, doing business in Bentonville is a media company. And I want to share with you what we've been working on for the future. And if you look at the statistics up there, we've have now 86, over 86,000 viewers and listeners. We have over 4,000 followers and 515,000 impressions. We've been extremely busy over the last, thank you. Thank you very much. Over the last year or so, we've had hundreds of podcasts. And with this journey, we've had some co-hosts working alongside of us making this happen. There's a lot of people working, the podcast video studio located across just the lawn here. This is a phenomenal uh facility. Great people working there, helping and supporting this. Also, there's some of our podcast hosts here today. I don't want to miss anyone, but I see Scott Benedict here. Scott is uh doing a great job as a host, and he's got his own uh uh uh moment or minute that you do, and you're all over the area. Scott, so thank you for that. Deanna Baker, where are you? Deanna, she's a co-host. Now, Deanna's History's Walmart, Scott's Walmart and Sims. And uh if you're we'd love to have you join Deanna on a podcast. Um I would love for you to join me on a podcast. I think I've seen uh Stafford somewhere. Where are you? Uh are you here? Um I thought I seen Josh come in at some point. Oh, there you are, Josh. Hey, welcome. Josh has a great show. Uh, and of course, uh, if you need a haircut, he can take care of that too, right, Josh? So, you know, um, anyway, it's it's been such a great team effort. And I just want to know, I can tell you how much we appreciate this live event. Now, a bit more about the future. About uh, we will continue doing live events. In fact, we're going to announce today a uh next upcoming live event. It will be in January. We'll tell you a bit about that in a few minutes. Uh, then we're also going to be working on other uh live events. Last week, I had the opportunity as a Walmart officer alumni to attend the event with Doug McMillan and with uh and then John Ferner. As you all you all know the news, what's going on at Walmart. And as I was sitting there and listening to Doug and John talk about the future of Walmart, the future of retail, and I was taking notes as fast as I could, Deanna, on my phone, trying to capture everything they were saying, because I was thinking about this event and future events, and I thought about something Sam Walden said. And because, you know, we were talking about Mr. Sam at that event, and you know, I just put up, there's lots of quotes that Sam says, but Sam says, ignore conventional wisdom and go the other way. AI is changed gonna change retail. It's already been a tremendous amount of change around omni-channel, and we're focused on that at doing business in Bentonville. That's our mission and that's our vision. But AI is the future change of retail. And then Sam also says is everyone, if everyone else is doing it one way, you can often find your niche by going the opposite direction. One of the things we want to do at doing business in Bendville is be your partner. We want us to come along aside each other, and we want to do several things. We want to help you and your company and your business. We want to walk beside you with change. Uh, we want to walk beside you and encourage you through our workshops that we'll be having at Doing Business in Bentonville. We'll be doing other live events, and I want to encourage you to reach out to me in a few minutes. You'll have you'll see my email up there, and you can reach out to me, and we're going to establish another group steering committee, Wayne, like we had in the past, for us to learn together for future uh live events and other things. Then Doug McMillan said this. You read this quote, but he stated it again last week when we were at the home office. Doug said, it's very clear that AI is going to change literally every job. You know, I began thinking about that after last week, and I was talking to Eric. I said, one of the things we need to do is have an AI, retail AI with people. Because my background, I spent the last 10 years of my career at Walmart as running human resources or the people division for Walmart. And we had a lot of change because we were moving from super centers and growth and from the original Walmart to super centers. We did all that. It was about change of roles and responsibilities. AI is going to have a huge effect on how we staff our stores, how we try to staff our companies, how we train our people and prepare them for the future. Doug also said, maybe there's a job in the world that AI won't change, but he says, but I haven't thought of it. So AI is going to change our future, but your partner that's going to be with you through this is doing business in Bentonville. And you're going to hear a lot about our future. Okay, so that's a bit about where we were. This is where we're going. So next thing I want to do is bring up my co-host. And uh Chase is going to come up, and Chase Benny, he is uh CEO of Retail Wire. He's become a good friend. We've had an opportunity to be together. He's going to uh to introduce our sponsors, and we'll go through that. Then we're going to introduce our panel and then we're going to get forward with our panel. They're they're they're such a great quality group. You've got to really love what they're going to do. But before Chase takes over, I want to thank Acosta and um where is okay, Troy. Thank you so much. Tony, Tony, thank you so much. Thank you for finding this venue. Tony, please tell Don we really appreciate this also. Pass it. And I I love this venue so much, but I need a little help. Maybe you could help me encourage Tony and Don. We could like come back, maybe, you know, to this. You know, you guys think about it. Okay. All right, okay. All right. Uh Chase. Well, let's welcome Chase. So glad you're being our coach today.
SPEAKER_02:Uh thanks, Andy. And you know, this kind of event wouldn't be possible with the sponsors. So um, yeah, Acosta, thank you so much for having us at the venue. It's very cozy, feels like a big living room. So um, and then uh, of course, RetailWire, uh, which I lead and represent, and Sev is here for my team. Um, a few of us here in Bentonville, and a few in New York and London, and uh, we'll be at NRF and some of the the next upcoming retail conferences. Uh Crisp is a very special company here that has a presence here in Bentonville. And um we will be bringing up Heather for a minute, uh, who's in town. And I hear they're having their own Christmas party. Uh so it's a festive time of year. Um Crisp is a great platform. I learned about them uh actually through Are, the founder, and um they just launched a new uh a new office here at the Ledger. So uh it's a very cool platform, all about clean data, which we all we're gonna talk about that too, um, in the world of AI, dirty data or clean data. Um and then Trive is another platform. Uh a couple of guys are here. Is Jason in the room? I mean in the back. I saw his I saw a hand go up. Um, another really cool platform that helps uh CPG suppliers communicate uh with uh the retail teams in a in a really seamless way. Um CSA, Cameron Smith, Hubbard, and podcast videos. And so with that, I want to bring up Heather from Crisp to introduce our panelists.
SPEAKER_00:Thank you very much. Um thanks everyone for being here today. As Chase mentioned, my name's Heather Merton, and I'm with Crisp. If you're new to Crisp, um, we are an AI-powered retail analytics platform focused on um helping brands and retailers better serve their customers by unlocking the power of the retailer-direct granular data across the business. Um we're headquartered here in Northwest Arkansas. We've got offices at The Ledger, as Chase said, and then also in Springdale. Um I'm super excited that you've enjoyed us or you've joined us today for this panel, AI and Retail. Um, I've I've been on a couple of panels and I've been applying AI into the category management space for about seven years. And one of the quotes I love is AI will not replace people. People with AI will replace people without AI, right? So learning, staying up to date on where AI is going, where it is today, and how you can apply it to your business is really important. Um, on our panel today, I have the distinct pleasure to introduce a handful of folks. Um, first, we've got Bill Aikens. I didn't confirm with you how to pronounce your last name, so I hope it's Akins, from the University of Arkansas. Bill is a purpose-driven executive educator and innovator whose career bridges enterprise transformation, applied AI, and human-centered leadership. Over 25 years, Bill has led digital and cultural transformation across the consumer goods, retail, and education sectors. Next, we have Chris Brintley, who joined, who leads digital consumer experience for L'Oreal's Walmart business and has nearly 15 years in e-commerce and CPG. He's held senior digital and e-commerce roles at the Bountiful Company, Kellogg's, Colgate Palm Olive, and Clorox. Jack Rudelik from Slalem works with companies across the central region to build strategic roadmaps that drive real business value, whether that's AI-powered transformation, large-scale product development, or internal tools that improve operations. And last but not least, Brandon Viveros brings over 15 years of experience driving digital innovation for leading CPG brands. As a seasoned agency executive and educator, he now leads the charge in redefining retail media at adfury.ai, helping brands unlock creative performance at scale. Please join me in welcoming our panel.
SPEAKER_02:So I've led uh quite a few roundtables and discussions like this. And one of the things, just to frame it up, is you know, we've got some experts here, and I'm here also to learn. So I'm also a student of this, and you know, a lot of us in the room, we want to learn about this stuff too. But the word expert is kind of a loaded term when it comes to, you know, uh new technologies. Uh, but show of hands, how many people here feel like they're an expert in AI? We got one, one brave soul who would say that. Um, so most of us we would we would say we're beginners. Um and so when we dive into this, that's the framework, open mind and really learning from those who are spending their time uh to be here and share with us. So, first of all, to kick us off, really the big question is in a world where agents are shopping, not just humans are shopping, there are agents shopping. What ensures that your brand will get chosen? So, on the topic of AI agents and agentic shopping, Bill, do you want to kick us off with your thoughts on it? Sure.
SPEAKER_06:So, yeah, I mean we've we've seen the news, right? So Walmart has partnered with OpenAI for Chat GPT. And to me, that is a that's not a novelty. That's a signal. All right, so that's that is a signal for uh for the role that AI is going to play in this very, very important space moving forward. So uh so I think it's important in order to answer the question, not only where we're going, but I think it's important to understand a little bit where we've where we've been and why this is such a pivotal moment in time for this. So if you kind of look at what has made brands successful in the world of retail, right? You have all these different evolutions of we started with uh uh uh PG coined the phrase the first moment of truth, right? It was all about the package and it was about that presentation on shelf, uh, which moved into the second moment of truth. Of then it was how are customers using your products and what was their experience and what kind of product reviews were out there. And typically most of us, you know, you'll you'll put a product review in when you've either had a really good experience or a really bad experience. You kind of miss that that middle of the road piece. But uh, but then it was Google deciding, saying, no, no, no, no, no. It's really about search and search and dising. I think there was even a book about it. It says that it was the zero moment of truth, right? And uh, and then that was where you're gonna win in retail. And where we're at today, though. And I think what makes this just such a such an exciting time in this space is um you kind of have to call this like the perpetual moment of truth, where when someone is inspired, your brand wins when you uh when when you are that when you are that gateway to make a frictionless experience and a frictionless purchase. That can be a service, that could be a product. But um, but but obviously in this perpetual moment of truth, uh, you know, frictionless commerce is is um is is where we're gonna win in the space to go about that. So uh when I when I teach at the U of A, uh we teach uh uh it's really applied AI. I am I am certainly not an expert in the algorithms and the machine learning underbelly behind it, but a lot of the NBA students were focusing on how do we actually apply this. It's been the hardest classes to teach because by the time I finish teaching something, it's obsolete and uh and and it changes. You know, sometimes as a professor, you can not that any professor ever reuses their content. That would never, never happen. But uh, but in this space, you just can't because it's it's changed so quickly. But in this, in this world of perpetual moment of truth, there's been a couple of key milestones. I mean, probably all of us remember the first prompt we did in Chat GPT, right? And you're like, oh my goodness, this is this is a game changer. This is really, really real. Um, you know, I get we we we saw um, you know, the first time we did an image and all those things. But there truly was this milestone moment for my students and I over the last three months, and it was this AI browser piece. Anybody played with Comet or Atlas or really, really heard of this? So, so so two major, major course, you know, AI, uh AI companies, OpenAI, uh produced this browser called Atlas and Perplexity, another AI tool, uh produces this uh this tool called Comet. And yeah, and and on the surface, you might look at this and say, well, the last thing we need are more browsers, right? You know, I already use Safari, I use Chrome. But what is so significant about this moment, though, is that to have an AI-infused browser, it looks like Chrome. It looks like Safari, it looks like Edge, smells like it. But it has the same exact interface about that. But what is happening now is that it's agent-driven so that I can literally be on a page on Walmart.com. Or I can literally, even before I even open that page, say, I am looking for noise-canceling headphones between$125 and$150. I need it to arrive in my house by Friday before my flight leaves. Uh, go. And it literally, you can it can take control of the browser and um it could literally go through all the iterations on that. So we're gonna talk about this. I think a lot as a panel today is what does that mean, right? And what does that mean for you as content? But We are at an influx moment. This is as important as when ChatGPT really made AI available to the rest of us. I mean, AI has been around since the 1950s, but it was that moment of Chat GPT that actually made it available to the non-data scientists, to the non-PHDs, you know, in statistics that can actually utilize it and understand that. So starting point. But uh, but but but I truly think we're at this influx point of now agents being able to do my shopping for me, um, it introduces a lot of opportunity. It introduces a lot of risk, it introduces a lot of things around governance that we have to think about as companies. But um uh I kind of go back to early DBB days when we were talking about uh e-commerce, right? And it was this uh, you know, getting, we were trying to get customers over that fear of putting their credit card on the internet. Like, oh, I would never shop online and put my put my credit card on the internet. That's too dangerous. And so so the dangers we're facing today, I mean, obviously, are going to be mitigated over time and about that. And so it is important for for brands that that we kind of understand now kind of what what that means.
SPEAKER_02:Well, it sounds kind of dangerous to let a browser just do some shopping for me.
SPEAKER_03:Yeah, I think, you know, just to you know build on, you know, uh what Bill was saying, you know, if you kind of think about you know the word agent itself, right? You know, if you think of an agent, sports agent, is somebody doing something on your behalf, right? And so we've gotten to this place with technology where there's enough data signals, enough compute processing power, enough data speed now that all of this can can really happen real time. And so we're looking at beyond just this ability to just find something, it's actually taking action upon what it is finding. And it's looking for signals, right? So these these agents, these AIs, these LLMs, they're looking for signals. And where do those signals come from? Well, they actually come from very varied sources, right? They come from your website, they come from the retailer's website, they come from your product page, right? They come from your retail media. And so now retail media, in the in the world of retail media, which is an area that you know I work in quite a bit, and we're working to make the process of creating retail media ads very, very fast, very quick, very efficient, scalable. We're really trying to create, you know, different ads for different target audiences. Well, those ads now become signals. And so how people will resonate with your message, how people will resonate across multiple surfaces with your brand, with your product are signals that are going to be very, very important in this uh future of shopping. And so that's where I, you know, I think, you know, think about what is it really going to take to win in this space, right? And is there, you know, is there an opportunity to win? Like as it was said before, maybe by the time we get done with this panel, the things that we talk about might be different, might have changed, right? But I it's really about consistency. Like I think of two words in this space. When I think about AI and think about shopping, it's two things it's trust and consistency. And there's trust that has to be played on two sides the trust between the AI and the human. Because when you look for something, you want to feel good about what you're getting back in return. That's why when you interact with AI in the non-shopping sense, you always get a response. And you always, you know, you always get the real happy, positive, you know, engagement, right? Because they want you to trust it. But then there's also trust between the AI and in this case, the retailer or brand or product, right? There has to be trust. And so it's looking for signals, it's looking for data. Where does that come from? Well, it comes from your attributes for your PDPs, making sure that's consistent. If there is a if there is a credential or a certification or a claim that is verified, and that claim is verified across other sources, like let's say consumer reports. Like all of those signals add up. And again, going back to all the compute power we have today, the data speeds that are available, the reasoning that AI has now brought to the technology. All of those things now come to play. So now when you are looking for something, you're not just simply looking for a bottle of laundry detergent, right? You are asking for much more than that. How to plan for a tailgate, how to plan for a party. I mean, we can even take it back to the laundry scenario and even think about like people may ask like how to get out certain types of stains or you know, um, you know, how to freshen their clothing. And you know, you want to be recommended. It comes back to trust, right? And then consistency. Being consistent in your communication. Now, I don't mean literally consistent, it's more semantics. Semantically consistent from platform or to platform or surface to surface about your brands or products are going to help you win because again, it comes back to that word, trust. The the AI has to have trust in that information in order to serve that back to the consumer so they feel good about it. And then from there on the purchase, how the purchase process works. Obviously, there's a lot of you know, a lot of opportunity through Apple Pay, PayPal, a lot of, you know, a lot of different ways to purchase. But really, it's about you've got to feel good about it. You know, I do think that there's still a lot of opportunity to really figure out what is the right modality for engaging with AI for the shopping experience. And it's probably going to be a multimodal type of experience where data is coming in through audio, visual, your browser, and all of that. And that's again, that's the great thing about AI, it has the capability of processing that across, you know, basically in real time. So, I mean, that's that's my, you know, where I feel like we're we're going with shopping and a lot of obviously a lot of different angles, but I still feel like it comes down to trust and consistency. And the more that you can provide that as a brand and the retailer as a facilitator of that is going to ensure that your product is well known.
SPEAKER_02:Uh, you mentioned trust and having your data accurate. And I was thinking about just how much of our effort is going to be going to um optimizing product pages for an AI agent versus optimizing for the human. Um, it's just a question that's come up. You know, it's like, how much of our world is going to be really marketing to AI?
SPEAKER_03:Yeah, it's it's a tough one right now because there are no like, there's no playbook, right? Because not only are your consumers using AI to find products, they're also still going to Google, right? And they're still searching the old way, we'll call it the old way through key through keyword search, right? And so what might work for Google through search may not work for AI, and vice versa. So you're trying right now, the the margin is relatively slim between those two and trying to optimize your content. So there is a lot of work that needs to be done in order to optimize that. I will say, though, on the retailer side with your products, you know, as many of your attributes that you can make sure are consistently completed, all of your claims are going to help in both situations. So I would definitely start with attributes. There is some fine-tuning of the content on your PDP in order to work for both models, uh, both through traditional search and through agentic search as well. You know, I was we were just even how I was having a conversation with a brand the other day, and they were talking about, well, they went ahead and they modified a lot of their primary hero images in order to work for, you know, for AI. But then Google was penalizing them for when they were doing this, when people when consumers are doing search for the same products. It's going to take a lot of test and learn. I would even encourage you to, you know, go into whether it's open AI or whatever LLM of choice that you choose to use and look for your products, see how they show up, and continue to test that day after day after day. That is truly how you're going to learn. And so just set up consistency there. And you can obviously do that agentically. Maybe you set up an agent that will do that for you on a consistent basis and kick out a report or let you know when that agent is run on on your on your products and try, you know, do a test and learn, do A-B testing, a lot of those types of things. Because again, the playbook hasn't been written yet. So it's hard for me to come over and say, do X, Y, and Z, and you will get success every single time because what I tell you today, again, may be done or may be different by the time we complete this talk today.
SPEAKER_02:So Chris, what are you seeing on the brand side of things?
SPEAKER_04:Yeah, so one, I both both panelists here, I think, kind of set it up well for us. And and somebody that works on uh a brand on the Walmart business, it's it's a question that I can tell you we're we're trying to have a conversation about on a weekly basis, right? So I think since Walmart announced the Chat GPT partnership, I've probably had no less than 100 conversations in the period of a month around like how L'Oreal is gonna get to a point to where we're uh a leader in this space or at par with our peer set, right? And so I understand, and as you all understand, working for brands, these are these are questions that we need to answer, things that we need to be prepared for. I think BV said it, well, it's a it's a bit of a we, we're we're on the on-ramp, right? And and on that on-ramp, we're gonna discover a lot of things, right? Are we gonna be able to accelerate to a point to where we're gonna meet our peer set at the level in which they are? Or are we going to be able to accelerate enough to exceed that, right? And ultimately as a brand, you want to exceed that. Uh, you want to be in a position where you're leading your particular category or or your particular item set uh for this particular retailer. And I think ultimately as a brand, we have to think about how do we come along with Walmart and also lead them at the same time, right? So I don't know if, you know, what sort of questions all of the different suppliers and manufacturers are having with Walmart. What I know is that they're still trying to figure that out, right? The moment I ask my buyer what they think about agentic AI and ChatGPT partnership and how we need to be prepared for it, uh, you know, a lot of him and hon, right? A lot of like, uh, well, you need to look at it, let us know if you need any help. Maybe if we learn anything, we'll let you know, right? So kind of typical, typical response there. And so I think for us, it's about uh as a brand, how do we become better educated around how we need to prepare for this space? So that's why, you know, I love to have the panel up here and learn, uh, learn from the expertise. And then I also think it's about how do we bring expertise to Walmart to help them understand what the future could look like for their platform. The good thing is, or I think all of us have noticed, right, over the last six months to a year, Walmart's made a lot of changes in the back end, right? Think about how often your CQS score has changed over the last year and a half, two years, right? I feel like we've 98, 45, 55, 75, maybe 10, you know, who knows what they're gonna do week after week, right? But that that shows that they're trying to figure out algorithmically how we're gonna change the way our system looks at these items. Um, and and also force us as manufacturers to make sure we're providing that information. So if I could encourage the group here when we think about AI, and and again, I think BB said it well, it's like make sure all your attribution is filled out 100%, right? I don't know how many of you know this. Moving forward, like we've gotten a little bit of a peek into what the new item creation process is gonna look for Walmart. You know, in the past, we've been able to get away with, let's just call it placeholder information to get an item number right. We know that, you know, Walmart requires that 18, 22 weeks out. Generally, we we don't have the content available or the attribution available for Walmart at that time. Walmart's gonna require that at item creation. So, what does that mean? That means that your marketing teams, your brand teams are gonna be, they're gonna have to have that attribution, that content ready for you at the time of item creation in the future. Um, and if, you know, we all know the the kind of pitfalls of not being ready for that, right? Uh, features late, you know, maybe our item set is not where it needs to be. Maybe we don't get to the mod in time. And, you know, we know Walmart doesn't like that. So I, you know, I would leave it and say that as long as we're making sure we're focusing on the fundamentals, where we continue to learn and understand how Walmart is on this path. And, you know, ultimately we push our internal teams to be engaged as well. We should be in a pretty good spot.
SPEAKER_01:Great. Can I add something to that? Um, just we talk, you've been talking about like the Agent Tech Shopper. And I feel like a lot of the companies that I talk with, there's this big barrier to entry of we're so far away from that. Um, first you have to do a data transformation, then you have to do an AI transformation, then maybe a workforce transformation, and then everything you're transforming to, like you said, it's obsolete because everything's changed. And you don't have to do that. I think there's opportunity, like, how can we scope it? How can we scope the products that we're trying to sell, what we have, focus on the attributions for a specific product set? How can we focus on where our data is best, where the supporting infrastructure is best, so that we can test and learn on how we can move forward with AI. You don't have to rebuild your company, rebuild your IT to get to this, we want to start doing AI. There's absolutely opportunity to see where, assess what you have, and start. And I think that a lot of people are afraid to just start because they hear a lot of their competitors say, we're doing this whole thing, we're gonna be with NAGENTEX shopping. Every company is different. Some companies are big, some companies are small, some companies have bad data, some companies have a lot of good data. Um, but understand where you want to focus and scope your initiative there. There's a lot of opportunity there, and I don't think that enough people realize that you don't have to just do this gigantic overhaul.
SPEAKER_05:Okay, thank you. Great discussion so far. Okay, our Bill's going to lead us off on our next discussion around strategic implications for your business. So, Bill, lead us off and we'll talk about the supplier group as we do that.
SPEAKER_06:Yeah, I mean, I think, I mean, as we as we journey down this new road, you know, I heard a great quote from uh from um from a guest speaker came to one of our classes, said that, you know, conversation is the new code, right? We do not have to be uh, you know, PhD trained, uh, you know, to to get to get started in this space to baby steps, right? Crawl, walk, run, uh, as we look at this. And so I mean, I mean, the importance of teams understanding just the basics of prompting, what that looks like. It's really interesting. You know, here we are, 30 years into standard search engine optimization and utilizing tools like Google. And still most of us aren't really utilizing Google to its fullest extent. And we would all claim, well, yeah, of course I know how to do a Google search, but a lot of us don't understand the Boolean operators and, you know, and actually how to get a really powerful search in. And unfortunately, we're not gonna have 30 years in in the world of AI to adjust into that. So, you know, understanding the basics of prompting and understanding um, you know, this is uh that I'm I'm establishing uh, you know, context in a role, I'm giving it a task, and I'm giving it constraints. Uh, and if I and if I can understand that, but I can, but I can do it conversationally. That's why I love that quote saying conversation is the new code. That's what makes it available to all of us. And so again, as we as we start to go down this road, I would challenge you to to really think of this in three simple layers as as you look at your product pages, as you, as, as, as you look at um, you know, your your uh your window into your your your product world and what that is. And and those three layers are you have the user-facing piece, which which that's what we're all familiar with today, right? This is what this is what our PDPs are, and uh, you know, everything that we're developing uh you know around that we've we've spent the last decade uh developing around it and enhancing that content. But I think these these these two other areas are are the areas that really challenge you in, say that we have our user-facing layer, but then there's this bot facing layer, and that's uh you know, that there's this layer of content now um that really has to be a little more AI friendly. AI loves bullet points, AI loves tables, AI loves for you to be very, very, very detailed on that. There's, as I said earlier, there's there's our double-edged sword, though, and that uh and that's you know, AI will do something called hallucinate that I'm sure some of you have come across. Uh, you know, AI, uh it it AI wants to please us, and I don't mean that in a creepy way, but but but but is but it but is that you know AI wants to wants to answer the question. And and in fact, uh it is penalized the way that these models are trained, it's penalized um that if um it's penalized if it if it says I don't know. Now, there are ways around that, and we could we could talk about prompt engineering all morning, and then we could talk about how you set up your uh your your tools and memory layers, you know, saying if you don't know the answer to something, I don't want you to hallucinate. I want you to actually prompt me for other questions around that. But uh, but but there was but there's there's there's there's ways to mitigate that that risk. And so I wouldn't I wouldn't have that be a uh you know a deal breaker for for for thinking about uh for for what what those layers are going to look like. So so you've got this kind of you know bot facing layer, and then uh and then you really have this kind of invisible layer underneath of uh the operational layer of instead of looking at um you know your your price comparisons, the model comparisons, your your your your comp says, all those pieces around so looking at that weekly to the point is that it is going to have to become a daily task. And uh and and the beauty about utilizing these agents, and and a lot of times, uh little of the way you define agent, but that's uh that it's a it it's a it's it's it is a subset of something we're already doing that just has an extreme amount of expertise. I mean, my business card, the first five years I lived here many years ago when I first started working with with Walmart and the Splur Committee, it's a retail analyst. Now so proud of that. I was an Excel jockey. I was not a retail analyst. And so now uh now I have tools that can take away that mundane part. And yeah, and we can truly do uh and and and and we've talked about this for probably two decades at doing business of Bentonville sessions of let's move the 80-20 rule, right? Of 80% prep the data, 20% analyze the data, where we truly have the tools now that we can flip that script around. And yeah, it should be 20% prep and 80% true analytics on that. So exciting time in that space.
SPEAKER_05:Good B V. You wanna you wanna move us and talk about the the new frontier?
SPEAKER_03:Yeah, so you know, I think as we've all kind of hit on on it at at various points, like you know, the new new frontier really, you know, comes down to you know how how we are engaging with technology in this in this new way. And I and I love you know what Bill said about just the way that you can interface with AI today. And then I've always called it natural language programming. That's what I've always thought of it as. Like I, as a kid in the 80s, I used to, you know, do some programming in my room on my TI-99 for a computer using BASIC, and I'd spend all afternoons just to, you know, make a very simple game, right? And so, but it requires syntax, right? And it was, you know, you can make a mistake and it wouldn't run. And there was a debugging, right? But now when we engage with AI, we can just, you know, whether you follow a prompt structure or not, you can give it any question you want and you're going to get a response, right? And so it, you know, you it you just ask it. And as I said earlier, I think, you know, chat to a certain degree, I think it is a limited modality, right? Could because it doesn't always see everything around you. It's not, you know, sitting down and thinking through, okay, how should I structure this prompt in order to get what I truly want to get out of it? Right. There's some friction there, right? And I think, you know, as we even think about agentic shopping as well, or or AI of AI-based shopping, what what what modality, what process is going to have the least amount of engagement based upon what you are trying to look for? If you're looking for a roll of toilet paper, I probably don't think you're going to go into Chat GPT and ask it for the best rolls of toilet paper. Probably not. I doubt it. You probably are going to either go to Walmart.com or go to Google and, you know, and add it to your card or add to your list, right? So there are going to be these different moments that you will engage in AI in different ways. You know, obviously there's been, you know, uh Google Glass is supposed to come out next year. You know, uh Meta has had some success with the Ray-Ban glasses, right? There's a new modality. There's been a lot of of pendants and wrist and bracelets. And so all of this uh ambient data that's being that's coming in so that you can just continue to uh add more data, more information, prompt it however you want, and you have that context, right? Because that's really the two things that really comes down to with why we are where we are with AI, is because you know, we have data, which we've always had. We've had AI, you know, which is we've had in various forms. Now we have this whole process of bringing it together with a reasoning layer, right? That's the part that really is bringing all this stuff together in conjunction with the compute power of the things I've mentioned before. So it's really this orchestration between those three things that really brings AI and the way that you engage with AI and how you engage with the I with AI will continue to will continue to evolve. Um, and I think, you know, there's that gives us a lot of opportunity and probably puts a lot of fear in everybody's hearts because we all want to make sure that our brands uh succeed, right? And so there's a lot of that fear of the unknown. Very quick applicable point to piggyback off of that.
SPEAKER_06:Game changer for me was uh utilizing the Gemini AI tool, but using the camera mode, uh, literally standing uh, you know, in front of the serial aisle at uh at Walmart. I have a son who um, I mean, he's definitely allergic to peanuts. And literally as if I was taking a picture, I held this camera up and looked at the serial aisle and um and I just used my voice. Just it was wasn't anything kind of fancy prompt. Me as an end-user shopper, and I said, uh, you know, what uh what what cereals are safe for my son? Uh given if there's if there's a peanut allergy behind that. And it was about 80 to 90% accurate, uh, where it literally, it was almost like a heat map and it and it lit up the product boxes uh for me that uh that it was literally going back and reading the the ingredients. And so imagine if uh if I was a cereal manufacturer that uh yes, it's in my ingredient list, but if I had gone into my contents on my PDPs and said, you know, no peanuts, you know, about that to go do it, that uh could even have pushed that higher up in the model to go do that. But it brings up the point, there was still human in the loop. You will hear this phrase in in the world of AI where as a parent, yes, it gave me a starting point uh from a shopper standpoint to be able to know where where to start. I'm like, all right, well, at least I need to isolate it to these two or three, but I'm still gonna read the box as a parent. Uh I'm not gonna completely go go trust that. And uh there was a human in the loop factor on that. But but I think of that again from a from a supplier side, uh, what needs to be developed on the content piece uh to ensure that um that those models become even more accurate.
SPEAKER_02:That kind of brings us to the topic of GEO. And so to set the stage on this, SEO is a term that means search engine optimization. And it's the science of getting your product or brand showing up in essentially was Google, was really the only one. And there's Bing and and uh DuckDuckGo and all the others. But the professional world of SEO has been focused on Google for the most part. Um and when we look at how you've already stated a lot of use cases for using generative engines or answer engines, there's a few terms like AEO, AIO, and GEO. And we've not decided what's the official one. Um, but I'm sitting here as someone who grew up with Google, and I feel like all my Google prompting skills are out the window. Um, I got really good at searching Google. And then ChatGPT came along and people started using that to write resumes and do different, different things. And I think we're at a spot now where the stats are there. It's majority, it's it's a huge majority of people have been using Chat GPT and other uh AI engines to find products, to do their shopping, to uh you know, find out which which products are safe and have peanuts in them or not. And it's it's we're taking that that with us everywhere now in different contexts. Um so when it comes to really getting your brand found in and cited in those answers, there's a scramble in the industry to try to figure out how to game that system and get your product to the top. Um, you know, but uh BV, I want to pass it to you and and talk a bit about that as brands are here thinking about, you know, we we we got really good at SEO and we were to the top of Google, but now the number one spot on Google is halfway down the page because the AI overviews are there and consumers are starting to use chat GP chat chat GPT and Perplexity and the others to find products.
SPEAKER_03:Yeah, and I think that's that's why I think Google has actually a great opportunity to win here, right? Because Google already has the mind share. Chrome has an exorbitant number of downloads, everybody uses Chrome. You know, I think I think the few for me in the future, what I see for the future, honestly, is AI-enabled search. And and I have this phrase in my head, and I don't know maybe how controversial this might be for this panel, but we'll know that AI has made it when we stop talking about AI. It just becomes the technology that powers the things that we use every single day. We don't even think about it. And I I think, you know, I think that's exactly how this will end up going. But obviously AI is a bud, a buzzword. That's why we're having a panel with AI in the title of the of the panel discussion. And and I think, you know, it's it's as some of the things I've said previously, it's it's going to take testing and learning and experimentation, and you've got to spend time doing that or dedicating resources to, you know, be that that research RD team, that test and learn team to continue to evolve. Um, and I think that's that's really where we're going to, where we're going to end up. Jack, do you know if you do you have any thoughts on that?
SPEAKER_04:Yeah, I'll just I'll just add a point um from a Walmart brand perspective. So um I agree, right? Like we we are on this kind of evolution, right? If you think it from a consumer standpoint, think about how you interact with with the internet as a whole today, right? Like I'll be honest with you. Um, you know, I spend some time in in in in LLMs and and I prompt certain things. And we have an engine uh specifically for uh at work that that is behind that kind of firewall, if you will. So we're not sharing data with the world, right? But um, you know, I think every consumer is on a different journey, right? I I want to say 10 years ago, I was showing my mom, you know, where to put the URL in on the on the web page, right? So, you know, we're all in a different set of journeys. So um, as I kind of alluded to earlier and and kind of what we've talked about here on the panel is that, you know, Walmart's gonna come along as as the consumer dictates what needs to happen on that platform, right? And so as we sit today, we're still in an 80%-ish SEO situation on Walmart.com, on SamSClub.com. Uh so that doesn't mean that we need to go and everything needs to be GO tomorrow. Uh, we need to be considerate of that, considerate of the behavior, considerate that most individuals who search those platforms search them in a keyword fashion. So uh, you know, all the things that you've heard over the last 10 or 15 years around uh make sure you're on the top of the keywords that matter the most, make sure you have the keywords embedded within your content. That is still important. Um, not to say, though, that we shouldn't be in this test and learn environment. And again, be on this on-ramp with Walmart. Understand what the needs are and how the platform is changing because it is changing. It's changed tremendously over the last 15 years. And I started working in digital on Walmart uh right before they purchased Jet. So I've been around for a while, right? We've seen this evolution come along. Um, I think from a content perspective, it's delivering the basics, right? What's Walmart asking for from me? And am I delivering it? And am I delivering at a level that is beyond what the metrics say I should deliver, right? So the water level in terms of the expectation from Walmart, or what we could or not get in trouble for is hey, you have to deliver X. We need to deliver above that, right? We need to be ready for that change. And so I would say encourage your internal brand teams, internal marketing teams to make sure that whenever that project innovation comes or you have legacy items, maybe where the data has not been aggregated to a level where it needs to be. Let's make sure we get caught up in that fashion. Let's make sure we be considerate of where the consumer is today, how they're searching the platform, how the algorithm is deciding on where your product's going to be ranked. But let's also live in this test and learn environment and be ready for these changing dynamics with uh platforms like Perplexity and ChatGPT and GEO and AEO. Like we should be considering what our product detail pages should look like five years from now. Um, not to say that needs to be implemented today, but let's go back and let's look at consumer sentiment. Let's understand what consumers are searching on the site. Let's understand what the trends are because ultimately we know that search behavior is moving in that direction.
SPEAKER_03:Yeah, and I think what's great about that is you have all that data. Once you have that data, you can synthesize it, right? And then come up with strategies in order to affect your, your, your opportunities on your P with your PDPs.
SPEAKER_06:I think that's outstanding advice when you look at uh when you kind of look at the the the whole model approach is that I think it's a it it would be a healthy mindset to take back to your teams to say, yes, today we're still competing for shelf space. And I mean that both in the physical space and and and in digital shelf space, but uh but over the next five years, will we not just be competing for shelf space, but we'll be competing for model attention and attention from these tools um to understand my product more so prompting is another piece and and we are mindful of time because we want to have time for for QA as well.
SPEAKER_02:Um but when we are looking at from the operational standpoint, um you have a whole team and you tell them we're all gonna use AI, we're gonna be so much more efficient. How do you actually, you know, what are some of those challenges that uh companies will face just from an operational standpoint? Um, Chris, you mentioned having having your LLM on a private, you know, it's not leaking your information out whenever your team is searching. Uh so uh Bill, do you have any thoughts on on this, just the challenges from that, from that side of things?
SPEAKER_06:I mean, it's a it's a new language, right? But it but it's a but this is but this is an easy language. This isn't one that we have to go uh you know and study as if we're interpreting a foreign language. Uh, I mean, to B V's point, it's uh that's why it's called natural language processing. Um I mean, I think it's it's being able to uh being able to have the skill set to know how to ask the right questions. But I would also encourage you to think about how do you reverse engineer things. So a lot of times we'll spend a lot of times in in in class saying, um, you know, we'll we'll take an image and say, okay, let's actually have the AI tool uh reverse engineer that image so that we can recreate it again. And I would challenge you to think about how do we reverse engineer how customers are interacting and talking to our pages and just using real language about that and understanding that.
SPEAKER_03:Yeah, I mean absolutely, yeah, absolutely love that. I think, you know, starting with the end in mind and working your way back is not natural for us to do, right? We always kind of start from the from the starting point and kind of try to get to the outcome. And I think as you as you work more with AI and you get used to prompting, like once you get to that end result, ask it, hey, how can I prompt you to get to this end result again quicker and then save it, right? You should have a prompt database that you share, um, whether you share it for yourself for your own personal reasons, or you share it with your team so that everybody can learn from that and have like that shortcut. Um, that's why I still think that there's some opportunity around the way that we engage with AI, has some, some, some opportunity. Keeping prompts memorized and a whole database of prompts doesn't feel natural to me, even though obviously I do it because it's a necessity.
SPEAKER_06:You want to know one of the best exercises to do for yourself in the world of AI as you're learning. Uh, the the the we literally do this on the first night of class for the MBA students, is we go in and uh everybody has their LinkedIn profiles, and we go into uh Cloud Perplexity Co-Pilot and uh ChatGPT and say, roast me. Roast my LinkedIn profile. You've got to have very, very thick skin. I wanted to cry like the first time I saw it. I mean, it was like, how insecure is this guy? How many do how many degrees does he have to have? I mean, it goes on and on. I mean, but but it was one of those I learned what the model was looking at in order to pull that back. And I mean, one of the best exercises you can do is uh roast your roast yourself, but uh uh, but but to go back and and and and roast your your your your your product page, right? And and uh you know, be able to go ask it uh and ask it from a from a competitive standpoint. And again, if you haven't played with the idea of agents, you know, utilizing just at its at its simplest form and being able to form uh you know just a small agent inside of Chat GPT of uh create one for your buyers, create one as your buyer, like a fake buyer and say my buyer is looking at one, two, three, four, and five on a daily basis, on a weekly basis, at our at our line reviews, everything else is a part of that. And and then ask that that quote buyer questions uh about about that content.
SPEAKER_01:Just want to add one of the I think one of the most exciting things to think about in this AI in prompting is how prompting's going to change. Like right now, it's all about like, and you kind of reference this, Bill. It's like, how do I construct a question with through text to get the answer that I'm looking for? And that's this point in time, is how we're interfacing with that. We're on the brink of, okay, eventually people aren't gonna want to like, people are gonna be relieved from having to text through their phone. They're going to have glasses that are understanding visual cognition and going to be uh able to take the voice at the same time. And that's going to be the new prompt. And in addition to those prompts of like, now there's that dual modality of like being able to take uh vision and audio and then be able to process that and return that into something that is meaningful to you. But also a lot of prompting is building that contextual memory bank of like outside of this question about what we know about like you and a little bit. And at this point in time, it's you're able to build a repo of what is the context of like who this person is. What I know about Bill, and I his kids are like eight years old. I'm making up stuff like he's eight years old. Uh, the last time this happened was this time ago. They didn't have epipens and this amount of amount of time. And there's all this context you can add to that. And the prompt becomes less about just the way that you ask a question and have a template that goes along with it. But everything that you can add in context into the memory bank to the LLMs and the systems that you're using to make those prompts so much more meaningful and such higher um confidence in that what it's returning to you is something that's meaningful and powerful to you.
SPEAKER_02:I got a final question, but Jack, I'll start with you. Um, as we uh, you know, we get inspired from this type of an event and we spend an hour thinking about this stuff. What's one thing that uh our our listeners can do when they get back to the office, or even just what should they be learning on the weekend in their spare time? Jack.
SPEAKER_01:Yeah, I love this question. One of my something that I think you should think about is what's holding me back. I think that a lot of us have these limited beliefs of well, I can't do it because of this, or we're not gonna be able to get there because we have these constraints. Kind of ignore the constraints that you think that you have, that your business has. Obviously, like don't be risk adverse. Um, don't just start putting like all of your company's data into an LLM. But think about like there's opportunity for us. Maybe we have to scope it down, maybe it's not gonna be as powerful as our competitors are right now. But if I want to run a marathon, then I have to start with running a 5K or running one mile. Start and remove the we can't get there because of this. Think about what's holding you back.
SPEAKER_02:Great. That's great. Chris.
SPEAKER_04:Yeah, I mean, simply I would say it's okay to ask questions. Right. It's uh it's okay to say, are we, why not? Um, what are we doing? You know, I work at a large, complex company. I know some of you work at large companies, medium-sized, smaller companies. You might be the copywriter in your office, you might be the one that that vets the content, you might be the one that thinks about how we're going to be prepared for for AI and and what we're talking about here today. Um, but I I generally practice keep asking until someone tells you no or tells you to stop, right? And, you know, in a large, complex company, there's a lot of moving parts, there's a lot of different individuals working on different things. And, you know, one one thing I would say that's helped us. Us is just bring it up in conversation, right? Listen, we have very complex things to do. We have a business to run. We have goals and things we need to achieve by the end of the year. I get that. You know, there are certain things that operationally we need to get better about. I get that. But what are we thinking about a year, two years, three years from now? What are we doing now to be prepared for that? So keep asking questions, keep asking why, why not? And uh, you know, ultimately I think you're going to force the conversation.
SPEAKER_05:Okay. What do you think about this group? Didn't they do a great job? Um thank you, panel. Thank you so much. Uh, we could be here another hour, and we have that many more questions for you. We could do that. But um there's a couple of things I mentioned on the front end that we're gonna we'll be doing to partner with you. Um one, we will continue our AI retail coming in uh in 2000 uh 2026, right? Um, we're gonna be working on uh AI with people. I mentioned earlier that uh that was uh my root role at Walmart, CHRO. And so we're going to be working on that. That'll be coming up, so watch for that. You can even sign up for our email alerts. I mentioned also about signing up for our email at uh alerts at doing business in Bedonville. Uh every every week, we will give you great information. We'll be a great source for you. As I mentioned, we're we are working really hard to become a media company and we're getting there. You've seen the statistics, it will continue to grow. Also, I mentioned earlier is that in the the ODBB, one of the most valuable things I had to help me guide the organization was a steering committee. If you're interested in uh being on that committee, you can reach me at my email and tell me you're interested. I reach out to I'll reach out to you. What we'll do, we'll have some virtual meetings along the way. I'll ask your suggestions and opinions on guests we should have here, topics we should address. Just like this panel, we got to keep asking questions, right? And that's what we'll keep doing to help you and grow in your business. Uh, the last thing we have is this January 29th, on Thursday, January 29th, our next event, AI in retail on media. So you can go ahead and sign up for that. You will notice the people in this room are very exclusive group. We'll maintain that. The people here or people like all of you working with Walmart and Sam's and this space. We're gonna keep it that way. We're gonna have a lot of integrity in this group. So I recommend before we uh uh cap the audience that you sign up for that event on January 29th. Thank you, thank you, thank you for being here. It's been a pleasure getting to see you. Thank you again. I hope to see you at our next event. Check us out, subscribe uh for our news email. And we want to thank again Acosta for allowing us to be here in this event. You thank you, have a great day.