Alright. Take it away, gentlemen. Great. Thank you so much. So thanks everybody for joining today. I hope you’ve been enjoying the the sessions. I know we have the consensus with the Market Logic crew there in the back has been that we need to get out to more of these sessions, both in the room but also, out there in the the in the central hall and hopefully in a maybe a evening session you would have had yesterday. There’s so many great conversations going on here. So in that spirit, we wanted to, together with Hormel, do a session all around their AI, adoption. So quickly introductions before we get going. My name is Joseph Rini. I’m the director of product management at Market Logic. I’m joined by Dave Anderko. I’ll turn it over to him in a second to introduce introduce himself, introduce Hormel, and then back to me quickly, thirty seconds on what Market Logic does, and then we really wanna open it up, put the slides away, and go into a discussion together about Hormel’s Alright. Thanks thanks, Joe. So I’m Dave Anderko. I’m the insights lead at Hormel Foods. I’ve been at Hormel about fifteen years, and I also have, some experience on the supplier side. So Hormel Foods, hopefully you guys have heard of us. We have a lot of brands. Maybe you might be more familiar with, like, Bacon, Chili, Spam, but we do have quite a few others like Planters, Skippy, Applegate, etcetera. So we have quite a large number of brands and, categories these days. So Market Logic, you may or may not know us. We are, the leading insights management platform. We’re used by of the largest Fortune five hundred companies in the world across consumer goods, pharma and healthcare, automotive, retail, and we provide an AI enabled, but really a human in the loop platform that these global organizations run their insights platforms out of. And then very quickly on a couple of our offerings, this is just part of our suite of offerings, but it’s two of the pieces that Hormel has started to adopt, so we want to just quickly cover them. We have our DeepSights Explore piece and our Personas. Both of these offerings very different but they’re set on us containing our customers’ proprietary data that they can then work with, run analysis on top of in the personas case, really speak to synthetic customers and so on. So, yeah, hope we can dig into those, here in our conversation. So the first question for you, Dave, would be just like what are you guys trying to achieve in bringing AI into your insights stack at Hormel? Sure. Alright. So start off with what is the purpose of insights at Hormel Foods? The way I like to view and I like to tell my team, our goal is to help the company make better decisions. That’s what we’re here for. Right? We’re gonna bring the consumer perspective and hopefully that’s going to lead to a better decision on what products we’re gonna launch or what different choices we might be making. So when I think very broadly about what I want AI to do, I wanted to do two things. Right? I wanted to help us make better decisions, and then I want us to help us make those decisions faster. So then still, get more into details as we go into this, but still keeping it at broad level, like, can we be doing how can we move faster? Well, that’s where you have AI as like a partner or an assistant. So that’s something that’s gonna be helping my team out. So it could help figure out, you know, how can I best run a study? You How can it help, you know, challenge some of my thinking? How can it help me write a questionnaire or review a questionnaire draft or summarize information for me? So how can my team use it as a partner to move faster and smarter? And then the other big unlock for AI is the amount of information that it can bring in, and really diverse information of that. So if you think of like, let’s just take a category for example, like bacon. So if you think like, here’s all the knowledge about bacon in like a big pie chart. So now, I I work for Hormel, I like to eat bacon, I probably know more than the average person about it. But I still only know kind of a small sliver of information about it compared to all the information that’s out there. So what AI enables us to do is go to all those different data sources and it can return it to us much faster. So then if we have more information and we have a better idea of what we’re doing, it should enable us to make better decisions faster for Hormel. So then like thinking in terms of actually how your roles would change given the the tools you’ve laid out there and what it can do for you. How do you see yourself and also your team shift? Yeah. So I’m glad we’re talking about this because I I have seen a lot of presentations as well you guys probably have as well. I guess you’re not AI ed out as you did come to this, thankfully. But sometimes I think people aren’t thinking broad enough about what the actual change is. So if you think about the old way of doing research, it’s you’re you’re working with, you know, your cross functional teams, generally marketing, you have to figure out like, well, what is the business objective here? What am I trying to do? What is the information I need? And in the past, the next step is usually secondary information. Right? I’m gonna go search for some things. But if you think five, ten years ago, right, that was hard to do. Right? You you had to go search through maybe find one old report, maybe find one industry type report and you read through that a little bit, you might get a little bit of information. But generally, it’s only gonna get you a little bit of the way. Right? And then you’re gonna have to move into custom research, usually a bigger study, and then at the end you bring it all together. Alright. Here’s what we found. Here’s what we’re gonna do. So what AI does is it really changes that second piece. Right? So the first part, right, you’re still going be working with your cross functional teams. What are we trying to do here? But now as I’m looking for secondary information, think of all the different sources I can have. Right? Through something like a DeepSights, now I have all my internal information. So let’s say, I I don’t know, maybe I’m working with innovation on trying to figure out what are some pain points with pepperoni. So now it can search dozens of reports that I have on pepperoni. Maybe it can search, you know, dozens of industry reports on pizza and pizza toppings and other things to do with pepperoni. And maybe if I have a social listening tool or I look at reviews on different websites, it can bring in so much more information and do it so much faster. Now, instead of only getting maybe five percent of the way there, I might be fifty percent of the way there, maybe further. So then, now, I don’t necessarily need to do a big piece of custom research. I may not need to do research at all or maybe I just need to do a quick like agile piece of research just to validate what I found. So it really changes how we’re approaching in the middle of going to of where we’re going as far as gonna and I think it’s important to think here is it isn’t just one piece on its own. It’s kind of that collective AI ecosystem and you’re thinking about all of those pieces joining together to give you an answer that you wouldn’t have been able to get to before. And so when I think about what that may mean for my team, Joe, is, you know, in the past a lot of expertise needed to be on running research methodologies, understanding what to run. Now that’s kind of transitioning to I need you to be better at understanding information and synergizing it. Because a lot of you may have had this some of the junior insights managers a lot of time if you ask for say, hey, you know, summarize what we have on pepperoni. I want you to use a couple of different sources. What you’re gonna find is right, they’ll do, oh, here’s one slide what I found from DeepSights, one slide I found from Copilot, one slide that I found from, you know, a Mintel report. Okay. That that’s nice, but what does that tell me? Like, what is the story that that’s bringing together and how are you weighing the different pieces of information, especially if you’re gonna use a copilot or a chat GPT? Like, did it actually come from real information or did it just kinda, you know, make it up from something where you’re not really sure where? So it’s gonna be understanding, you know, what are all the data sources, what’s a quality data source, and then putting that together into a story. So like what does that so what does that save for this old idea of depth versus speed? Straight off there. So based on what I said, you might think that I’d say, yeah, you can get really high quality information really fast. Right. But there’s a huge caveat with this, Joe. Here’s the thing. So let’s let’s play out kind of the scenario I went through to where you’re finding all this AI information and you’re not doing as much actual custom information. In a couple of years, where’s all the consumer information coming from? Right? I’m gonna it’s gonna run dry. So what’s really important is that you the base of a lot of your AI stuff needs to be high quality information. So what I’m doing right now with my team is figuring out what is all the foundational research we need to make sure that we have a deep understanding of consumers that allows us to to kinda skip a lot more of that situational type research. Because right now at Hormel, I’d say most of the research we do is situational. Right? There’s something we need to test a design, innovation, whatever it is, you know, a product change. But if we have a great base of foundational insights and that’s got to be a good mix of quantitative and qualitative to get that deep understanding, then that is what allows us to move faster. So you kind of have to move slow in the beginning to then be able to move fast. Yeah. Makes sense. Makes sense. Let’s shift gears a bit to a little more of a tactical question. So like how were you guys going about thinking through how you’d evaluate AI solution as you decided to look for one? What kind of key outcomes were you considering and so on? Yeah. So great question. So going back to what I said earlier on is I want us to move faster and kind of smarter and better. So for the faster piece is it’s got to be easy. You guys are probably all aware of the consumer frictionless trend. You know there was a good presentation this morning at was at the Heineken woman who was like, you know, coworkers are people too. Right? My team are people as well. They want it to be frictionless. Right? Like, they don’t want it to be difficult. If I’m saying this is gonna make your life easier, but it’s really hard to it’s really hard to log into, it’s hard to figure out, it’s hard to tie it back, like it has to be very easy to use. The next piece of it is it has to be trusted. I mean, there’s a million different AI companies out there these days. How do we know which ones that we’re gonna trust? And as I’m thinking of stakeholders who are gonna get the information, you know, how do I know which ones they’re gonna trust? Because as of right now, just even with consumers, right, if if you go into and you’re presenting to a brand team and you have a result that they don’t like, what’s the first question they have? Where did you find these people? Right? Like, where did these people come from that hate my brilliant idea? So now imagine, now we’re replacing real people with potentially an AI source, whether it’s a synthetic persona, synthetic data, or just an AI summary of things. Right? There’s even more barriers than to believability. But if it’s coming from like a trusted company or a trusted partner that we’re working with, right, it’s gonna be a little bit more believable on the back end. And then the last thing, you know, what is the ROI on this? Not necessarily needing an exact dollar amount, but getting back to, right, is this actually helping us move faster? Is it actually helping us make better decisions? Are we able to, you know, eliminate a tool that maybe we’re paying for? Are we able to do, you know, less situational research? Right? It’s gonna have some kind of benefit in the long term for us. Maybe just quickly on like sort of what was your experience with DeepSights of course, but then maybe also for the audience would be interesting to know how important was it for you to go for a vertical specific offering versus a copilot or generic AI tool? Yeah. DeepSights, houses so it’s actually we through Market Logic, we call it the hive. It houses all of our research. So that’s our research management system. All our projects go in there. All the reports are in there. Then we also have feeds and uploaded from other sources like Mintel, etcetera in there as well. So what’s great about that is that we know the source of the information where it’s coming from. Because like I said earlier, if it’s coming from something like Copilot, who knows where it’s being, where it’s coming out of. So as I think it was worth evaluating DeepSights, right, the first thing is is this accurate information? And because it’s our own trusted information, right, it is coming back pretty accurate. And then we have to figure out, okay, you know, is this actually useful? Because a lot of times if you’re going to copilot to try to figure things out or chat GPT, right, can get you sometimes it can get you some good information, sometimes it can’t. I think we found pretty quickly that it returns pretty strong information. So going back to like the pepperoni example, if you type in there, like, what are some jobs to be done for pepperoni? What are consumer pain points for pepperoni? It does an excellent job going through, searching it, coming back, and and returning helpful information for us. And it’s, you know, not just for the insights team. Right? This needs to be broader. So we brought in, like, the marketing team to test it. We brought in, like, the category management and sales because we we know they would be a big user, and we had to make sure each of those was able to find some useful information in there. So maybe just to you, Joe, as you built out DeepSights, I guess, you go through a little bit more of, like, what was your philosophy in building it out? Yeah. I think, like so our approach, especially vis a vis some of these more standard, the Copilots and the standard LLM tools is of course, as you’ve been stressing, right? It has to be based on the customer’s proprietary data. That’s point number one. It’s not about going to the large language model to find information. It’s leveraging the positives of the large language model, it can express information, it can synthesize information, but always being sure, as you’re saying, to feed it quality information that only your organization would have and that you and the insights team are even vetting. Then same with the personas, I think we’ll get to in a second. The idea, the philosophy is, again, these tools can be built in standard in house operations, but it’s about ensuring that they really go to the customer’s data, reflect your understanding of your customers to your stakeholders, right? So those are two areas that we really try to focus upon. Maybe shifting gears to the personas offering. What was your thinking at Hormel in starting to experiment in that space? Maybe initial feedback there as well. The personas. Yeah. So this is actually building out those consumer personas that we can query. Kinda getting back to your AI the AI ecosystem. You have to say so if we’re summarizing old information, right, that’s great. But that’s only one piece. Right? Say if you’re using like social listening or reviews, right, that’s great, but that’s only one piece. What personas are to me is they’re a lot more dynamic. Right? It’s something that you can it’s actually based on a real person. I think that’s we’re finding out is we need to make sure it’s very based on a lot of data, especially qualitative information. So that you’re getting a truer response like you would from a consumer that you’re not necessarily going to get from summarizing information. As odd as it sounds, it’s it’s almost by building up a fake person, it almost allows you to get closer to real consumers Because you can I mean, it’s it’s like a real consumer that’s on demand right there? You can ask it whatever you wanted to ask it. You can ask it for feedback on concepts. You can ask for feedback on whatever it is. And what’s nice about that is, you know, instead of it just being, you know, a generic, custom GPT or something like that, no this is a deep layered consumer persona that’s going to give you much better and more articulate feedback. It’s interesting you say that because about half of the personas we set up is based on a customer’s existing segmentation So they’ve got these named personas and we bring them to life in the system. The other half is maybe we’re actually doing the analysis work on raw data, maybe transcripts, maybe UNA survey data. There it’s like we actually start to really learn and formulate and crystallize an understanding of the customer’s customer group that maybe they don’t even have because they haven’t done that exercise of surfacing it. So yeah, great. Maybe shifting gears to future ways of working and the one big topic that’s been in the air in this conference, but I guess just generally is the interplay between people, humans and AI and the human in the loop. Where do you see that going over the next couple of years? Yeah, that’s a good one. So how we talk about it, and I think we stole this from someone, so it may even be somebody in the room. We call it, you know, real intelligence and artificial intelligence with the real intelligence coming from people. So the way we I like to think about it, Joe, is, you know, AI needs to be more of that partner or assistant. Right? We need the real intelligence should be guiding the strategy, guiding what it’s looking into. Even if you’re having, say, an AI agent do all the research for you or you’re using, you know, AI persona to do it, but it has to be guided from the person and then AI needs to be the assistant. If you’re going to AI to say, okay, here’s the problem. What do I do? Or, you know, what what do you based on this information? Right? That’s not what we want. Right? AI should be your assistant. It should be your partner, but we need to make sure the person in the real intelligence is actually guiding the strategy and guiding the decisions. What about so we’ve talked a lot about how the insights itself works with AI tools. What about already or maybe what you foresee coming in terms of how you’re interacting with stakeholders internally and maybe even externally, suppliers and so on. How is that going to change likely with Yeah. So actually this is a so I’m gonna go on a little bit of a rant here, Joe. Because this is I’ve heard some different things about you know, is that we have all these AI tools and whether or not, you know, you need to allow the the whole organization to have these. So I see my role as I want Hormel to be a consumer first organization. And the only way I can do that is through democratizing insights. So some may push back and say, well, geez, if you let, you know, the sales team or the marketing or whoever it is have access to all the insights and it’s all summarized, well, aren’t they gonna use it the wrong way? Probably. Yeah. That’s that’s gonna happen. But you know what? That’s okay. Because if I think through where would I rather wanna error? Would I would I rather teams either not use insights or find something on chat GPT to use where God knows where that came from or just everything has to go through me and my team? Or I’d rather have them have this access to this awesome database of all of our best information. And sometimes if they misuse it, you know what, I’m okay with that, Joe. I don’t mind that. You know what, I’d much rather have an argument with somebody about the correct interpretation of an insight than whether or not they should be using insights. So to me, I feel like we need to expand it as broadly as we can. You know, whether it’s corporate communications trying to use it for PR or the sales team trying to build up a sales story, even like r and d that’s trying to build up, you know, trying to develop a new product and trying to figure out what consumers may want. So to me, say you don’t democratize the information. Put it out there. It’s worth the risk. Yeah, makes sense. So I guess that kind of builds towards where do you see this? What would you love to see in a year, two years in terms of how you at Hormel are using and the insights team are using AI within the organization? So so that it all goes back to making better and faster decisions. So I think the the key switch is gonna be is the skills that I need on my team in the future. So right now, a lot of insights managers, they know a lot about methodologies and they know a lot about the research piece. I think that’s gonna get lighter and lighter. Because I think we’re gonna be able to kinda off outsource that work to both our research partners and potentially like AI research agents that can do a lot of that. So what I need my team to be able to do is to really understand the business questions, to really work much more with the cross functional teams to understand, right, like we’re guiding like we’re guiding the consumer strategy in the right way. We can rely on other experts, AI and real to do more of the research. Then with all of this information, I need my team to be able to synthesize that and to tell a better story and to actually make more impact with consumer information in the long term. Sense. How about for you, Joe? Where do you see it going? So first of all, to echo what you said, mean, I’m in the product space. I guess maybe a foreign tech product space, a foreign area to many people in the room, but I’m interfacing with developers, with customer facing people. And just like you mentioned, that generalist push or movement is happening across knowledge work. So it’s not just an insight, it’s obviously that now you can use these tools to do much more generous approaches. I guess in terms of what Market Logic sees happening, I mean each year we come to a conference like this, we see these big shifts in what’s being talked about, what’s being accepted, right? A couple years ago it was will we use Gen AI or AI. I think that ship has sailed long and far and now it’s about how are humans more and more reinserted or almost with friction intentionally inserted into processes so that these systems are not just running wild if you will. I think that’s one exciting area that we at Market Logic are gonna keep building towards. Building out applications with our insights customers but in very targeted ways where we’re leveraging the skills of our insights customers and partners and like you said, letting the AI do some of the work, but ultimately keeping the humans strongly in the in the in the driving seat for the foreseeable future. Joe Rini, we’ve got maybe one quick quick quick question. I’m trying to see if I can squeeze it in. In a portfolio of dozens of brands, how do you track and identify those that are in most need of help support or additional research? So actually so a big change that happened at Hormel a few years ago for the for the best is we started a brand segmentation. This was a huge problem in the past. Biggest brands and the smallest brands all wanted everything from my team. But fortunately our leadership actually said, alright, now here’s the top eight brands. This is where the resources go. This is what you need to be doing. And so it’s actually made it a lot easier for us to say no. So if one of the smaller brands, I won’t mention which those are, if they come up and say like, I need research, we can say no. Like, you’re not a priority brand. You’re not going to get it. Now they’re still disappointed, but at least we feel empowered by leadership to do that as we’ve done the brand segmentation. Very important. Awesome. Thank you. Thank you both. Yep. Yep. Great. Thank you. Yep.
