Loading summary
Transcript72 lines
- [00:00]
A
Hi, this is Dave Glick, SVP of Enterprise Business Services at Walmart. I'm here with SRI on the CPG Guys Podcast.
- [00:09]
B
Hello and welcome to the CPG Guys Podcast. Set at the intersection of commerce and tech. Your hosts Sree Rajagopalan and Peter V S Bhan explore how brands and retailers engage consumers in a digitally driven world. And now, here are the CPG Guys.
- [00:29]
C
Hello and welcome to this episode of the CPG Guys Podcast. I'm of course Sree, your co host and also CRO and co founder of Think Blue Consulting, your trusted partner in your omnichannel development journey where you can get in touch with me at sri@thinkblueconsulting co not. Com. Please do listen to my older daughter's music@www.riaraj.com who now has a December tour five cities ranging from Chicago, DC, New York City. Of course we had to do LA and Toronto and follow Lara Raj. My younger daughter is a member of the world's fastest growing global girls group Catseye, now the proud winner of MTV VMA as well and will be on the road November 15th through mid December, Toronto, United States as well as Mexico City. Excited to be here recording with our guest today but unable to join me is my co host and co founder pvsp who also moonlights his head of Industry and client engagement at Flywheel, the commerce acceleration division of Omnicom. Make sure you're subscribing to our podcast on your preferred listening platform where you can get our latest episodes and go back to consume some of the 530plus episodes we've already published with brands, retailers and service providers to our lovely CPG industry and retail. You know that the CPG guys have a long established relationship with Walmart. Our number one and two highest downloaded episodes in our history of five and a half years is Walmart number one, Walmart number two, the world's most elite retailers, what I call them. So we're excited to have another area of the business represented. Today's episode probably one of the most profound one for today's times. Dave Glick joined Walmart in 2023 and serves as SVP of Enterprise Business Services leading what is effectively the IT department. His professional experience included 20 years nearly at Amazon. So please join me in welcoming to the podcast Dave Glick. Dave, how you doing? Welcome to the podcast. I'm thrilled that I can have a discussion with you on one of the most important topics hitting our industry today. Welcome.
- [02:24]
A
Thanks. Super excited to be here and you want to talk about AI is that the story?
- [02:29]
C
How'd you guess? Dave? Yeah, that's the word of the hour. While many call it, I call it artificial intelligence. How about that?
- [02:37]
A
Yeah, it's funny. I go back and visit friends and we talk about AI, then I come back to work and we talk about AI and so happy to do so here.
- [02:44]
C
It's definitely one of the most transformational trends of our time, so. But we're excited to have you on the podcast, of course. In the show Digital Notes of this episode, we'll include links to your LinkedIn profile and needless to say, of Walmart. So let me get it started right away. I think when I use the word. You are the SVP of enterprise services. What is your role? What does the day in the life of Dave Blick look like?
- [03:06]
A
Yeah, well, my organization spans all of the corporate functions. So finance, technology and services, people, technology and services, global governance, customer care, payments, fraud, directed spend, and it. As you said before, I like to tell people, we ring the registers, we pay the employees, the associates, and we close the books. So what else do you need to run a company?
- [03:29]
C
You know, fraud and shrink is something we're going to spend an hour just discussing on our own. Right. So likewise, AI is another one. We could spend a whole hour just talking about just the word, what it means and what it doesn't mean. But for the purpose of today, we're really going to take a deeper dive into the AI world. Right. But before we do that, though, tell us what a day in the life looks like. So Dave gets into work early in the morning and what does it go like? Are there crises? Like shrink alone has got to be a full time job.
- [03:56]
A
Yeah, I come in, I go to the cafeteria and get bacon and eggs every morning and a Diet Coke.
- [04:02]
C
I get a hazelnut coffee. That's my favorite.
- [04:06]
A
I'm sure mine is less healthy than yours. I have a few meetings. We check in on the big projects. What I really enjoy is sitting down and going deep with some of my team or some of the customers. We have status meetings and we have executive meetings and all those. But the things I really enjoy, we've got some of this written on the whiteboard behind me, but getting in and understanding how we can use technology to support operations. Then usually the end of the day, I go home. I just moved to Bentonville. I go home and walk around my driveway for about an hour to finish closing my rings on my watch. And I usually talk to folks in Seattle or on the west coast for my team. And so, you know, Most of the day in the office, but I enjoy getting outside as well.
- [04:50]
C
Are you the 10,000 steps guy? 8,000 steps guy? For me it's 10.
- [04:54]
A
I have to close the Red Wing. 630 calories, which usually ends up being about 10,004 to 5 miles.
- [05:00]
C
I've had a pretty slow start to the day. I'm only at 4, 24 steps, 0.19. I got a long way to go today, Dave, so send me some astral telepathy on closing the rings today.
- [05:11]
A
Yeah.
- [05:11]
C
Let's talk about as an early adopter of technology, you personally in your life and Walmart is leading the way for the industry as a retailer with adoption of technology. It's even in what your CEO Doug McMillan says about being tech as an anchor. Talk to us about the different aspects of business services. Who all do you touch? Who do you all. Who are your primary customers within Walmart?
- [05:33]
A
Yeah, you know, Doug has been great. I joined just over two and a half years ago and he's been talking about AI since the beginning. And Walmart has said we need to be at the forefront of AI. About six months ago, you know, we people were worried about hallucinations and biases. About six months ago, we said we just need to start building. And I sent off one of my principal engineers and said, you know, go build an agent. You know, one of the great things about my organization, as opposed to other tech organizations and other companies, is we have the services or operations teams right next to the tech team, all under my team. And so, you know, the finance services team, your customers, I'm sure will know we've got invoice to PO matching if the system doesn't do it. One thing we can talk about later is like claims and disputes process. That will be familiar to many of your listeners. But we have a lot of people, you know, either FTEs or contractors who spend their time following SOPs. And so we came up with this idea a few months ago that like anything, you know, we should build what's called an agent builder or I named an agent builder. Anything you can stick an SOP into and it spits out an agent to do that role. And you know, if you think about our business, you know, any operation really as time to serve, cost to serve and quality, you know, if we can have some of this work offloaded to the agents, the mundane, repetitive tasks, you know, that will decrease time to serve and cost to serve and presumably increase quality as well.
- [06:58]
C
No doubt about it, Dave. AI is transforming everything we do. But people confuse ML as AI, they talk about processing large volumes of data and getting insights from it very quickly. As AI, can you give us the difference between, especially wearing a retail hat, the difference between ML and AI?
- [07:16]
A
Yeah. My boss Cresh, talks about when he was in college, and I think it was maybe the 70s or 80s or I don't want to age him too much, but he was using neural nets and they called that machine learning. And machine learning was like pattern matching. Us humans can pattern match maybe two or three or four variables at a time. With machine learning and with the original AI, we could put hundreds of variables or thousands of variables in and suss out patterns. Often it was numbers more than letters. With Genai, we're creating something new. It wasn't just pattern matching. It was looking at, this is something that's never been created before, but based on my knowledge, we're going to create something new based on your prompt. And then, you know, that was like all the rage for about a year, right? And then, you know, we started with agentic AI now, and agentic AI is not, you know, you think of gen AI as chatbots and talking back and forth and having a conversation. Agentic AI is the machine taking actions. So today we have a lot of work that's being done both internally and by contractors who just follow SOPs and do sort of mundane things like resolving PO versus invoice mismatches. We came up with the idea several months ago of could we build an agent builder? And we've already written out all these SOPs. Could we stick the SOPs into the agent builder metaphorically, and have it spit out an agent? That's one of the things that we've been doing recently. We free up our people so they can do more interesting things while the machines take on the mundane tasks.
- [08:51]
C
So, Dave, there are three pieces of AI that I think of. There's the agentic one that you just address, there's the generative one, and then there's a predictive one. I like to call it gap. Generative, agentic, predictive agent. You covered what is generative AI and predictive AI?
- [09:09]
A
Yeah, Predictive AI was sort of the original machine learning, the original pattern matching. Hey, we've seen, we have all this data telling us what customers do or what vendors do. And, you know, how do we get smarter? Use that data to get smarter and putting the right products in front of vendors. And that was basically looking at the past and predicting the future. Now generative AI does some of that as well. We have huge, multi billion, tens of billions Parameters, rather than it used to be 100, 200 parameters that we're using in predictive AI. But that looks for what's the right next word. And it can, you know, the, the main use was like chatbots and assistants, like, how do we make the people more productive? And then as we get to agents, they're actually doing things right, they're taking actions on behalf of us. And so that's how I think about sort of pattern matching to assisting, to doing so.
- [10:04]
C
What was the strategic reason that Walmart decided, hey, we're going to develop a unified AI framework with four super agents instead of continuing to build standalone tools. Tell us what the four super agents are and then what was the strategic reasoning behind that?
- [10:18]
A
Yeah, well, let me start by telling, you know, many of us have intranets, right? And oftentimes it's very hard to find anything. In fact, on Walmart's intranet, you know, we came up with like short URLs, so that if you needed a credit card, you go to WMLink credit card and so on, but you have to know the secret code word. And so what we wanted to do is as we build tens and hundreds and thousands and tens of thousands of agents, we didn't want people have to go look and find those. And so the idea is the super agents are one place where you can go for your job and touch any of the agents that are backing those. And the four we have are built around Personas. So the first one is Sparky and that's what you interact with if you go to our app or to our websites for consumers. The second one is for developers. We call it Vibey, kind of a play on Vibe coding. And that's focused on the developers and developer tools and making us more efficient. The third one is Marty, and that's focused on suppliers. So you can go or you will be able to go and set up ad campaigns or other things in Marty. And finally, the one I think is most interesting, it turns out the one I own is the associate AI. And we're still working on that name. But the idea is you can address all of your work and all the things around your work life in one place. And so if you're in finance or in merchandising, say I wonder what my sales were yesterday. You can go to, go to our associate agent and ask it and it will find the right set of data and the right app that it brings up, the right UI that brings up, so you can research that and go deep. You know, our associates look for discount cards. And. But you know, one of the top things we got at our benefits help desk was I lost my discount card. So now we've built an agent which can replace that. And so you can go in and say, I've lost my discount card and we'll send you a new one or set up your direct deposit or you need to go match POS to invoices or we have questions in the store, an associate has questions in the store, they can look up a product. So all of those things you can come to one place so you don't have to remember in your head or in your bookmarks where to go.
- [12:26]
C
You know, Dave, what you're talking about is no joke. I mean, for a retailer of your scale, with 4700 approximate plus stores and the associates you have is one of the world's largest employers, trying to put this out for that many associates is no joke. So is this all work in progress? Is some of this being touched and used today? Like how? What's the vision here?
- [12:47]
A
Yeah, Sparky, you can go shop with Sparky today. The associate agent, we are releasing in the next few days our first application on that. And we have dozens of agents which will back into that over the next few months. And we'll be releasing as time goes by. And then our developer agent, Vibey, we're all using, we're all using that today. And so as one of my staff says, this is not vaporware, so that's good to hear.
- [13:12]
C
So one of the things the industry frequently Discusses is any AI agentIC, AI or AI based agent is only as good as the data input that comes in. So I gotta imagine you all have done some extensive work in the background which couldn't have been easy. So that when Sparky is used, it actually gives the right outcome. What's the process to make sure the data is what goes in is good in the first place?
- [13:39]
A
Yeah. You know, one, one way you can attack this problem is we have to make our data better, which ends up in failure every single time. But then you say we have this application and let us go run it and see, see where we see the data falling down. We can manually audit or use another LLM to audit that. But I always start from what action we're going to take or what problem are we going to try to solve and then work our way back to what data is right and wrong. And, you know, people really, you know, every single person who's done ML will tell you data cleansing is the hardest part. I think I'm sort of one step removed from actually hands on keyboard for this. And my feeling is like a lot of this data has been used for many years, many decades. Right. Our financial data. We're not changing the algorithm on the financial data, we're just making it a little more discoverable. And so what I expect is it'll be 90 something percent. Right. But what we'll start discovering is places this trip. Security too. Like where we have security through obscurity, we will find places where permissions are not right. And we need to go track those down quickly. And when we find data that's not we, we need to surface it and track it down quickly.
- [14:49]
C
And I gotta imagine behind all of this is individual LLMs. Like how does that work? Are there, is it one large LLM or. Each of the super agents is its own thing.
- [14:57]
A
Yeah. So we, we use models from all the hyperscalers we, we aren't training, or we haven't trained ours, our own yet, but we use whatever the right model is for the right thing. Sometimes we'll use two or three different models and see which one comes out correctly.
- [15:12]
C
Yeah, like LLMs. Right behind each of these is strong LLMs. I gotta imagine that you're continuously feeding, like how does that process work? Am I wrong when I say LLMs?
- [15:23]
A
Yeah. So we have these super agents which will reduce cognitive load on the users and help with discovery. But behind those we're building many what I'm calling nano agents. My colleague Hari came up with this term and I'm trying to get it to stick, but lots of little agents, and each one of those could use a different model, a different foundation model and different training. On top of that, we've built some boilerplate stuff, set up a gatherer agent, which we always use that to go look in this database, this database, in this database, get some information, and maybe you have a summarizer agent which summarizes all that, then an action agent, and then insert into a database agent. Those four agents all use the commercially available LLMs from the hyperscalers, but they're tuned in different ways and they're trained in different ways.
- [16:13]
C
Man, you got your work cut out for you. I mean, to make this feasible at the scale you're dealing with for that many associates, you're talking a form of agents is what my takeaway would be in this case. So let me remind our audience that today we're speaking with Dave Gleich, SVP Enterprise Business Services at Walmart, and we're getting deeper into how AI is serving associates at Walmart. And later on we'll get to how the vendor community can also find some benefit in the future. But what's the early feedback? You know, describe the practical feedback you're getting back from associates as they're using these super agents and all the agents that support the super agents. How's it improving their lives? Are they saying that's what it does?
- [16:54]
A
Yeah. So we actually did a hackathon. One of the things that I'm excited about is the engineers are building agents for our users. But even more exciting than that is the users building their own agents. One of my. I have a few. I speak a lot in sound bites these days. And one of the sound bites is bring the engineer close to the customer. And I've been trying to do that for decades, but, you know, when the engineer and the customer are the same person. So if the user can write things to build their own, that's the best way, because then you don't have this miscommunication of this is what I thought you were going to build and this is what you told me to build. We did a hackathon a few weeks ago which was all business users and we had 42 projects. I'll give you an example of this person's mornings are filled with. I take this spreadsheet and I take this spreadsheet and I spend a couple hours validating these. And so they were able to write an agent to do that in like 30 seconds. And so they can spend their time doing interesting things, interacting with other associates, driving transformation rather than comparing spreadsheets. And that's super powerful and they love it. On the other end of the spectrum, we built this thing called the digital accountant. You know, we have auditors, internal and external, and analysts who go look in different accounts and make sure they're there. And we do that sort of sparse inspection. But with a digital accountant, we can look at every single account and understand if there is anomalies. We showed that to our controller last week and all he said was, wow. And so, you know, he's not using the tool, but he sees the power in it.
- [18:24]
C
I'm going to say. I'm going to say double wow. One of the misconceptions I personally had, and I know a lot of people in this industry are going to have the same misconception, is that you have to be a coder and you got to know Python or some sort of programming to be able to build agents. I think you're dispelling the myth. Is that fair?
- [18:43]
A
It is True. And like six months ago, you would have had to be an engineer. We actually got together one night after work and we're like, we're going to do vibe coding. I want to build the agent builder in six hours. And so I brought a couple engineers and a couple non engineers with us. And one of them, one of the non engineers, she's like, well, I want to try. And so we installed our ide, our integrated development environment on her machine. And she like, that night went and built her own agent to do some accounting work. And every week she comes in and tells us what new agent she's built. And so I struggle because it's probably pretty complex for me, but she's just diving in and she's very curious. And, you know, this has led me to another sound bite, which is the future belongs to the curious. Like, you have to go push, push, push against it.
- [19:29]
C
There's no doubt about that. The future always belongs. The reason Walmart has been successful, you guys have always been leaders in what you do because of curiosity. There's no doubt about that one. So, so it is true that you don't have to be a super programmer to develop an agent. You just have to have the curiosity, some basic learning. So the CPG guys, we could build our own agents in that case, you.
- [19:52]
A
Could build your own agents for sure. And you know you can. We actually use a command line development environment, but you can use it, you know, you can go to any of the big hyperscaler and they've got their little box. You type in a prompt and it'll go, start building something. It's, it's quite amazing. And they, it's funny, the engineers. I have one engineer who's in Virginia, so he's an hour ahead. And so like roughly three days a week I'll wake up and have a text from him saying, you won't believe what happened. You won't believe what I did last night or this morning. And so, you know, that's one of the really fun parts. Both the business folks and the engineers are just have kind of got a new lease on life in terms of, here's all these more interesting things I can do that I couldn't do six months ago.
- [20:33]
C
Crazy impressive, Dave. You got your hands full for sure, but at the same time, you're getting outcomes. I learned technology or two about super agents. Let's talk about the people aspect for this, right, that word artificial agent, artificial intelligence and agentic AI is still relatively new with a lot of misconceptions. Some people get intimidated by it. But the most important thing to do is empower people, which you're doing by doing the hackathons and training people. You can do this yourself and bringing curiosity to the table, but you got to lead a large organization through what I would call a significant technology change. How do you actually do that and bring widespread adoption amongst associates?
- [21:18]
A
Yeah, I mean, I think Doug has had a great example for us. In every one on one, in every town hall and every global leadership meeting, he's talking about AI. We set a time to do that. And so I've tried to mimic that with repetition. And so one of the things we're doing. I used to think that engineering was the hard part of transformation. I was at a conference a year ago and the people before me were saying, hardest part is change management. And I was like, no, no, no, I'm an engineer. We know engineering is the hardest part. And over the last year I've realized that that's not true, that bringing people along is. And so we actually started something where I would go to a town hall of another VP or senior VP and bring one of the engineers with me. It was kind of like Iron Chef. We'd start the meeting and say, what do you want us to build? We went to the talent acquisition town hall and they said, build something that reads a resume and comes up with some pre screening questions. As we were doing a fireside chat, Lo and I, John was standing over to the side building an agent. After 10 minutes he showed us the UI had been built. And then after another, it's about a 25 minute thing. And by that, and he built the UI and he asked for some feature requests and he moved some buttons around, but it shows the art of the possible. And we did that first and then we said, well, that's the art of the possible, but it's not very accessible. Then we started bringing a non engineer with us and she was showing, here's something that I did. I installed our development environment. Here's something I did. And that was much more accessible. And every time I did one of these, I'd get two more invitations. And so, you know, that's, you know, 100, 200, 300 people at a time. But if you, you know, do those a couple times a week, you're touching a lot of people.
- [23:01]
C
Yeah, you empower people to be curious, but to take ownership and write a little bit of the destiny about owning the space and building the space, which is incredible. But with large tech organizations at Walmart, how do you keep that Alignment going around strategy, culture, and most important, being people, people led. Because there's a misconception that as agents come, the reliance on people led will go away, which I don't personally believe is true love, to get your opinion.
- [23:28]
A
Like, who knows what the future holds? But, like, I think the best thing we can do is empower our people to have them driving the change rather than change happening to them. You know, I had a boss who once said, you know, stress doesn't come from. From working hard. It comes from lack of control. And so, you know, if we can make people feel like they have agency and they're empowered and they're able to control the future, I think that's. That's what I think of when I think of people led.
- [23:55]
C
You know, there's no doubt about it, because most of us, we try to control what we have no control over, instead of controlling the controllables, which is what I believe causes a lot of stress in the first place. Now, you've been doing this for about two years at Walmart, as you said earlier today, and you are empowering people to kind of own their own destiny. Here with agentic AI, you're doing hackathons, et cetera. How's all the feedback been so far and participation been? Do people feel more empowered? What do they tell you?
- [24:22]
A
I think people are super excited about it. I mean, there is, you know, there's a baseline of, you know, change is hard, but broadly, the people, the people who dive in and are curious always come back at the end of the day or the next day or at the end of the week and say, thank you. Like, I'm having more fun at work than I've ever had before, and I feel empowered. And so we get a lot of positive feedback with the people who are being curious. When I started coding 30 years ago, maybe a little more closer to 40 years ago, it was about where you put the semicolon. Make sure you close every paren you open. Can you malloc in free memory? It's changed a lot. And so what we're going to look for, I think, in new engineers, is I. Are they curious? Are they tenacious? Are they gritty? I used to say the two most important things in work are persistence and resilience. And then HR funded a study over two or three years, and they came back with tenacity and grit. And I think those are exactly the same thing. Persistence and resilience and tenacity and grit. But I think those are going to become more and more comfortable because the Actual technical skills change and the cognitive load and the barrier to entry have come down.
- [25:34]
C
My first tryst with programming was Cobol 84. What was yours?
- [25:38]
A
Basic.
- [25:39]
C
Yeah, I learned BASIC first. That's not true. COBOL came second hand in hand. I actually, Dave, you brought back a memory of me going with my dad. He worked for this British company called Imperial Chemical Industries in Calcutta, India and he'd take us on a visit to the mainframe. We had to wear special booties to walk into that room. It was fully air conditioned back then when central AC was not the way for everything in India and offices. And one of the favorite things he would give me is leftover punch cards. And the way that mainframe would work is there were more negative results coming out of that, as in wasted punch cards, than they were positive ones. But yet they kept gritty. And the reason I gave this example is my dad was tenacious, gritty in ensuring that the mainframe was bought in to simplify life. And I remember these punch card incidents. My older brother and I would take these leftover punch cards and make cutouts out of them for playing when we were eight years old. That was my first twist with computers and computing power. I want to get into a fun one now which has to do with claims and disputes. I believe you're leaning transformation around claims and disputes. Very tough area, arguably lots of data, lots of disagreements across human capital over there. I believe you're going to move mountains here. So please, everyone's waiting to know, shed some light what's going on.
- [27:03]
A
Yeah, I was talking to folks, I said, you know, I'm going on this podcast, the CPG guys, I bet they're interested in claims. And you know, beforehand I said that to you and you started laughing.
- [27:12]
C
Dave, to me, let me declare just transparently, 30 years in CPG, that's some of the largest brands in the world. PepsiCo, J&J Revlon. Last role was CCO. General Mills. Claims and disputes was a way of life. So yes, very curious, those listening to the show are going to be as well.
- [27:29]
A
So we're basically started doing kaizens and looking at what are the causes of these claims, especially the ones that are disputed and we lose the dispute or.
- [27:36]
C
For our audience, tell them, what is kaizen?
- [27:38]
A
Yeah, a kaizen is. It's. It's from Toyota, the Japanese word for incremental improvement in steps.
- [27:45]
C
Right, I've heard of it, yep.
- [27:46]
A
Yeah, yeah. And so this is something that the Toyota folks lean. Six Sigma. This is one of the vocabulary. But you sit down For a couple days, everybody who's involved in a process and you write down the as is process and you actually go through and next to each step, decide if it adds value for the customer or not. And what you find is 90% of the steps don't. But having the people. The other term I'm going to use is Gemba. The people who live in the Gemba, which is the shop floor or the point of action. Having the people who work in the Gemba sit down with the engineers or sit down with the process folks and people actually know what's going on. Like, you don't want a bunch of VPs in the room. I often come for the readout, but I never go because I don't add any value. But, you know, I'll give you an example of something we found is that if, if we put we want three cases of two on the PO and on the invoice you said we sent you six, that's a claim. And that claim turns into a dispute. Hey, we noticed these prices don't match. Could you update your catalog, please? And this person went from eight figures, you know, in a year to four figures in a year of disputes or, excuse me, claims. And so this is the kind of thing that, you know, I don't know if it's the most sexy thing, but like, it gives me the dopamine hit. I really like getting into these things and, you know, that's real money back to the bottom line, both for the vendor and for Walmart. And I'll put in a plug if this resonates, please hit me up on LinkedIn and we'll give you a call and try to do a, you know, a Kaizen or a Kaizen.
- [29:20]
C
I'm going to tell you right now, claims and disputes. If you conquer this space for vendors, it's a game changer for the industry. So I'll answer the question right now. Should people be hitting you up and wanting to understand more and actually help you build a better product? Sounds like you've already made a lot of progress. It's a game changer for the CPG industry. So congratulations on actually not only focusing on sexy stuff, but basic operational stuff where huge productivity can be bought and operational improvements can be done. The most important reason to do that is people can. You know, in a CPG company, if you don't have to spend time doing this for hours and hours and hours and the amount of working capitalized, it can be reinvested in making better product at a better price for the customer. At the end of the day, you and I, our common master is the customer and we want to deliver for that person at the end of the day. So the more we spend on how we can obsess about the customer and these processes streamline, the better off all we'll all be. So that's what I mean by that. I'm going to guarantee you right now, CPG guys, guarantee. Take it with a pinch of salt if you want. This is a game changing moment if you pull this off. So for those listening, please, if you got an opinion, please do voice it. So I've heard the term stop being nice, start being kind. Why has that resonated so strongly with you and how does that translate to enterprise leadership?
- [30:44]
A
Yeah, what we find sometimes is if you have a culture of being nice and I think they call it Minnesota nice or something, but you don't actually have conversations.
- [30:54]
C
I may not think about that because my last role in corporate America was out of a Minnesota company.
- [31:00]
A
Yeah, you don't actually have the conversations you need to have or at least that people talking behind each other's back or you know, or just making, not, not making progress. And many of us, most of us, I think at Walmart or you know, big CPG companies are ambitious. We'd like to move forward. And when I was a young manager, I was scared of my employees, right. I didn't want to give them feedback, I didn't want to give them critical feedback. And I had one guy who was great, super smart engineer, he's actually now a cto, but you know, his attitude was such that I had to sit down and have a conversation with him and I was sure he was going to quit because he didn't like feedback. And we spent a half an hour and I had seven bullet points and I didn't know if he'd come back the next day. And then his roommate called me that night. So we were riding home and Ed couldn't stop talking about your conversation. He was so appreciative that you spent time giving him this feedback and you were honest with him and that changed this guy's career. And so one of the things I've learned is it's kind to address, give feedback and help people grow because it's kind of the most important thing you can do to help people grow.
- [32:08]
C
Feedback is a gift, Dave. So you also actively mentor people?
- [32:12]
A
I do, I have, I have sort of long term mentorships. I've probably got, you know, a dozen folks who I've known for 10 years who I'm super invested in their careers. I have a guy who I ran into on LinkedIn that I've talked to a couple times on the truck ride back from the new home office, I gave him a call and he's like, well, I'm stuck in my career. What should I do? And I give him three steps that he could do. And I said, or you could send me your resume and come work with us. You want to live in Bentonville? And he's like, oh, I'm open to Bentonville. So one of the good things about mentoring is it a. It increases your circle. And one of the things I think about is every interaction is either with a future customer, a future boss, a future employee, or a future referrer. And so this investment you make, while you could think of it as altruistic, it's also very practical.
- [33:09]
C
No, no. But what a fantastic framework you're giving for people. Every interaction treated with positivity, unless proven otherwise, and everyone benefits from it. The world becomes a better place when everyone benefits from it. It's that simple. At the end of the day, because you never know when you need a, when you meet a new person, what their objective is and what comes of it. Right? And if we shun away many of those, you and I would have never met today if we hadn't just decided to engage and talk about AI and what's going on with Walmart. Right? So that's as simple as it gets. Now, let me get to the last question for this discussion and this episode of, of the CPG guys. We call it Fast Forward. So I asked you to look forward. You know, originally we said three, five years. Three, five years of. The pace of change of technology is too crazy for me to ask you. And I don't think any of us can predict what's going to happen in three to five years. But let's look near term horizon, 6 months, 12 months, 18 months maybe tops. How do you see enterprise business services evolving with this whole world called AI?
- [34:11]
A
Yeah, we get this question frequently and my answer is always like, I don't look at things that are going to change. I look at things that are going to stay the same. Right. Our customers will always want everyday low prices. And for me, that means we have to deliver everyday low cost. And so we have a new tool set. And so if we've reduced by 50% the number of manual PO to invoice matches we need to do, and it took us two years to do that, how do we get the other 50% in three days or in three weeks. And I think, and again, I, I and many of my technical friends two years ago were like rolling our eyes at AI and saying, oh, we've seen RFID and we've seen blockchain and all these things that have sort of never really met up to their, their hype. But we've really found that this is actually not overhyped, it's underhyped. And so the things would take us a year to do can often be done in a week. And things that used to take us a month to do, we can do in hours. And so what I expect is, or certainly in my space, I haven't changed. I still want to have great time to serve, great cost to serve and great quality and serve our customers, be that as internal associates, customers or vendors. And so this is just another tool in my toolbox to do that.
- [35:25]
C
Tenacity and grid is what I hear, right? Commitment, commit, stay tenacious, stay gritty, Deliver what you commit to. Don't get carried away and try to predict what's going to happen. 18 months, control the controllables and basically let the world know that AI doesn't have to be this big scary word. It isn't going to dramatically kill everything, it isn't going to take over everything, just run with the flow. One of the most important things you said today is be curious and you don't need to be a crazy Python based programmer. For those who don't know Python is a programming language to really be in the world of agent AI, it sounds like anybody with curiosity can build an agent AI. Is that fair?
- [36:03]
A
I think that's completely right. And you see people on Twitter who are saying, oh, I built this over the weekend, I built a whole new application, I built a new ERP system over the weekend. You're probably not going to do that, but starting with things like comparing these two spreadsheets or every year I have to go through my email to find all my receipts for charitable donations. So can I ask, you know, the AI in my email to go, go get all those? It actually failed, by the way, but I'm going to go. It's probably I failed because I wasn't tenacious.
- [36:35]
C
Tenacity and grit.
- [36:36]
A
I called the sales guy for this product and I'm like, hey, this didn't work. And he said, oh, did you try different prompts? I'm like, no, A, I'm not being tenacious, but B, your product should not make me try these. And so that's what I'm trying to do. Is not ever have to say to my associate customers or my customer customers or my vendor customers, did you try this? Did you try that? It just works the first time.
- [36:58]
C
Perfect. Dave, what a fabulous conversation. I thoroughly enjoyed this kind of dispelling some of the myths, understanding how important this is, how this can be beneficial for associates, how to drive productivity and value the importance of tenacity and grit in this. Most importantly, be curious. Try to do this yourself. It doesn't require rocket science. Let me thank our audience for listening to this wonderful episode. Leave us a rating and review on Apple Podcast, Spotify or your favorite listening platform. It informs us how we're doing as well as if you're having the right conversations. To all of you listening in, thank you from Peter and me. You make the show happen to all our sponsors, whether this podcast, our parties, events, hosted dinners, having us at the panels. Thank you, thank you, thank you. The show doesn't exist without you. My big takeaway straight from Dave. You can do it. Be tenacious and gritty. Overcome AI by being curious. Dave, what a fabulous conversation. You're always welcome back on the CPG Guys to keep dispelling the myth that AI is some crazy rocket science. Thank you so much for making time for us.
- [38:00]
A
Thank you for having me. This was awesome.
- [38:02]
C
That's a wrap. We look forward to speaking with you on the next episode of the CPG Guys Podcast. Thank you.
- [38:21]
B
The content in this podcast episode is provided for general informational purposes only. By listening to our episode, you understand that no information contained in this episode should be construed as advice from CPGuys LLC where the individual author, hosts or guests.
- [38:36]
C
No.
- [38:36]
B
Nor is it intended to be a substitute for research on any subject matter. Reference to any specific product or entity does not constitute an endorsement or recommendation by CPGuys LLC. The views expressed by guests are their own and their appearance on the program does not imply an endorsement of them or any entity they represent. The views expressed by CPGuys LLC do not represent the views of their employers or the entity they represent. CPTGuys LLC expressly disclaims any and all liability or responsibility for any direct, indirect, incidental, special, consequential, or other damages arising out of any individual's use of, reference to, or inability to use this podcast or the information we present in this podcast.