TMRE@Home 2026: AI Personas and the Future of Consumer-Centric Innovation

So quickly about Market Logic, don’t wanna spend too much time on, on us, but we are, the market leading insights platform. We are used by some of the largest companies in the world to essentially run their insights teams, their global insights organizations out of our platform, and we kinda work across the consumer goods, pharma and health care, retail, automotive spaces. And you can, you know, take a look at some of the the logos on on the slide there. And what we offer is essentially an AI enabled platform that keeps humans in the loop, so keeps all of the insights professionals who are in our platforms on the customer side in the loop, leveraging the power of AI to make the most of all of their research, data, and insights. And today, I really wanna speak to you about our personas and synthetic respondent offering, which way to answering a lot of the, the pain points that, we are answering overall as a company, which is all around this customer research dilemma. Insights teams and their stakeholders need to act fast or they miss opportunities. They need to prioritize budget while making the most of investments, and they need to try and make as much as they can of their audience and customer intelligence, of course, without guesswork in as far as that’s possible. And it’s not just the insights teams. It’s really feeding that throughout the organization into marketing, r and d, and other business stakeholders. So the Persona’s offering, as I said, it’s one of our AI enabled solutions that keeps the insights or marketing professionals in the loop, and it essentially allows our customers to run forward looking audience intelligence. And I really love that term, and I’ll go in a little bit more into why we’re starting to use that more than just the persona’s offering as I talk about the full spectrum of what we’re now offering. To use this audience intelligence at scale without sacrificing research rigor. And our personas all synthetic respondent program is really based on current customer data. I think that’s a crucial piece that I’ll I’ll stress as we go through here. We build unique personas and panels based on our customers’ trusted data, providing really hyper personalized personas. And there’s a couple different flavors that I’m gonna talk about in a second to allow that really customer specific personalization. And as I’ve said, we are really now across the spectrum into both qual and quant audience research. So I’m gonna go in a demo in a second, but I really quickly want to take you through sort of the spectrum of what we provide in terms of this call and qual quant approach. Let’s start with qual on the left. So I’m also gonna cover the right hand side, the quant focus. But if we think of qual here on the left, this is really those qualitative LLM based chats that you’re likely familiar with that we offer. And there’s a couple different ways that we build these. I’m gonna demo them in a second as well for our customers, or I should say, also allow our customers to build the MIM platform. We set up existing personas that our customers would already have built based on segmentation work, usage and attitude survey data, other brainstorming internally, transcripts and interviews, voice of consumer. So a lot of times our customers will actually have persona programs set up internally, and what we’re really then doing is bringing those to life in a large language model based chat experience, which I’ll demo. But more and more, we’re seeing customers who also don’t have those personas built out already and or maybe they do, but they’re older, out of date, or they don’t cover the full spectrum of the customers that they want their stakeholders to be able to interact with. That’s where we also offer a couple different, what we call persona builder capabilities. And this allows our customers to really come to the system, prompt it for specific persona that, you know, probably doesn’t exist previous to them coming to the system, and then that’s spun up on the fly based on that repository repository of our customers’ knowledge. And we also have a flavor of that that can look to usage and attitude survey data and other types of structured data sources, really granular and high impact survey data that we can then enrich those personas that we can build on the system. On the flip side of that is our quant offering. In a second, I’ll show you that as well as I take you through some of the common flows that we’re seeing our customers start to use with these synthetic respondents. This is a newer, offering for us, but it’s now, being piloted or should really, say, used by multiple customers. This is, more of a quant focused offering, so less of a chat experience. It’s really coming to the system. You’re still interacting with synthetic personas, but you’re doing this at a large end level. So it’s not really a chat based experience. It’s like uploading content and getting this regimented feedback across potentially hundreds of personas in one go. And then we do a quantitative processing of all that content to give you back really robust statistical outputs, like top two box, you know, the distribution of the Likert responses. And, additionally, we let you go in and interact with those panels in a way that’s actually not really doable with humans. And here, I just have a quick quick look at what this looks like. So, right, you’re uploading concepts, you’re getting back these really robust scores. I’m now gonna show you, both of those as I just, take over the screen here. So I’ve now come into, our persona’s environment. And imagine I’m an inside professional business stakeholder, and I’ve come to the platform to interact with, some of the personas and start to ideate, think through some early, you know, maybe concepts or product ideas. So very typical for our customer base, there will be a number of potentially, you know, categories with prebuilt personas. As I mentioned, they might look like something like this. Right? So take a look at Martha Gomez. She’s, you know, of a certain age cut, f m g FMCG consumer, personal care. So she would have lived in a PowerPoint presentation, had been disseminated internally even prior to the existence of these Gen AI solutions, like our personas. And there was always this challenge of how do you bring this persona to life so that marketers, R and D stakeholders, insights, etcetera, can target campaigns, can target their thinking to these personas. So just, for the sake of time, I’m jump into an existing, chat that I had going. And, you know, imagine that I’d come into the system, grab Martha, and I simply wanna explore a little bit, her day to day, activities. And you can see I simply post that question and got back this nice, meaty response, quite tailored as to to what this particular synthetic persona does day to day. And then it was like, okay. Well, actually, I’m focusing maybe a new product idea somewhere in pain points or or challenges faced by this type of persona’s children. So what are pain points you’re facing in your children’s hair care needs? And I get the response back from Roger that. You know, it’s a constant battle and so on and so on. Right? So we can go a lot of different ways with this, and I’m gonna take you through a couple different journeys. We see our our insights and marketing professionals doing in the platform. Let’s just imagine that this was the initial stage that I, as the insights person, had come to the platform for today. And I’m gonna communicate this downstream to a team that’s ideating with me around this area. I’m simply going to say, okay. You know what? Let’s take an infographic of that chat, a little bit of an understanding of this persona so that can help us feed some ideation and further product work that we can do. And what now happens is everything we know about Martha, that chat that I just had, any other output she she had are sent away to the to the system, and it comes back with this great infographic. Really high impact. I think you can, you know, sort of just take a second to look at this, and you can appreciate that it’s captured what we talked about. You get this really visceral understanding or feeling for probably who Martha is, pain point she’s feeling, and then I can communicate that downstream internally and use that to keep ideating and keep the customer in the loop here. So that’s great. The next thing I wanted to show you, though, is, imagine that I’m now in the system, and I actually want to, talk to a persona that maybe doesn’t exist in the system. So I’m looking. K. My company has done some work in these couple categories I mentioned, but, actually, you know, I’m now here to ideate on a completely different space, and, I need to simply figure out a way to get a hold of a synthetic customer that I can ideate with. So what I’m now gonna do is open up this great persona area, and this is that second piece for this synthetic that I mentioned. There is no existing persona, but we’ve got this great repository of knowledge that we hold on behalf of our customers. Let’s comb that to create a persona. So I’m simply going to ask for a German consumer and hit go. And now the system is gonna start to comb the repository of content and look for all of the rich data that exists out there for this this request that I have, it’s gonna, in a second, pull back a description for me. So every essentially doing a lot of the work that traditionally insights innovation teams are going away and doing to create these personas, but not really enabling marketers, business stakeholders to come to the system and spin those personas up on the fly. So just, you know, just skip through that fast, it takes about forty five seconds to create that. You can see that, you know, I’ve gone away. I’ve created that persona. So I just changed screens here very quickly. And we can take a look at the details of what was looked at here in the repository in order to create this persona. And now I can go ahead and start chatting with them. Right? So a super targeted way to to create a persona and then speak with it. What I wanna demonstrate here, though, is our AI capabilities. So not only can I create these personas on the fly, but, hey? It’s a lot of work to actually speak with them. So I’m gonna start a new chat here, search for Jessica very quickly, and simply ask the AI moderator to understand this response, hair care needs and arrive at a product idea. And what’s gonna happen when I hit go here is the system is gonna reflect on that request that I have, and it’s gonna now carry out the conversation on my behalf. So it is you know, it’s first it’s formulated kind of a plan for how to pursue this conversation. And then quite quick, it’s gonna start asking questions, getting Jessica’s feedback, and then this is back and forth, right, happening. Right in front of me is the insights user. In fact, it’s it’s so fast that it’s hard to follow the eye. So you can see the quality of the questions being written is also quite an improvement on what we see humans typically doing when they’re interacting with these systems. So not only are we getting a kind of higher grade and but also a faster chat experience. The system’s gonna carry that out back and forth. Typically takes about three to five exchanges for this for it to conclude that it, you know, it’s arrived at the output that’s looking for here. And it’s now concluded that AI conversation. And if we just wait for a second, it’s gonna process that and load for me this, you know, summarization of what’s taking place so I can see what the goal of chat was, what’s been carried out, key findings that were surfaced. Of course, I can go back and read the whole chat, export it, and so on. But, actually, this is already quite a high impact output that I’m getting here. Quickly, before stepping over to the, quality quantitative focus, I wanted to show you another very, very common use case that we see, which is customers coming to our personas to get some qualitative feedback on new concepts and so on. So I’ve just created two fictional hair care concepts. Now let’s imagine I’m back in that chat experience. I’m simply gonna drag those concepts in and ask the persona, please. So very common use case. One or more personas is now gonna take a look at this image. They’re gonna analyze it. They’re gonna take a look at all of the concept, print, and copy and so on. And in a second, this particular persona, Jessica, is gonna give me back what she thinks about both concepts. And I could then go further, right, asking for further improvements, helping, to to rework the concept, and so on. And this is really a common use case that we’re seeing, this qualitative investigation of images happening in our system. So you can see here she’s taken, she’s preferred one to the other. She’s given some really good feedback on why. And as said, I could not take this in all sorts of different directions. So that’s great. That’s what we’ve seen customers using the personas, for to date. I’m gonna switch over to that quant focus output, which is something that’s fresh for us, and I think really I’ll drive home now the the the kind of qualitative difference that this quant piece is adding for us. So I’ve uploaded that very same concept here, But now I’m gonna go to a synthetic panel of personas that are spun up in the system. And just for illustration’s sake, I’m gonna ask twenty of them to weigh in here. Typically, this might be done at, you know, a hundred to two hundred personas. Let’s run the pipeline. And now what’s happening is each of these personas is being exposed to both concepts, and they are going to all weigh in on them. And then we’re gonna take that and quantify that exactly as we as we see customers doing when they’re using real respondent tests out there. And here we go with the scores. So, you know, both have gotten relatively low scores for this panel. I can then, you know, go and take a detailed look at what was said about each of the personas. And a nice little analysis I can run is I can simply ask a large language model. Hey. Take a look at those twenty responses per persona and point out what the key drivers were for the top two box. So that’s twenty six percent who were in the top two box, you know, what were reasons for rejection as well. So sort of a re addition of a qualitative piece on top of this quant analysis that I think also really differentiates how we’re letting teams do this type of investigation. And with that, I’d like to come back into the information and simply conclude by saying, you know, feel free to scan the QR code, get in contact with us if either the personas or the synthetic piece is something that you’d be interested in talking about. With that, I come back to you for questions. Thank you, Joe. That was fantastic. We do have time for maybe one or two questions. First question that came in, other than internal primary research data, what other data is being used to create these personas, and what are you doing to validate the output to ensure it’s reflective of the consumers? Yep. Great question. So, indeed, it’s it’s absolutely not just primary research. I think in that on the fly creation piece, saw where I prompted the system to create, you know, the German that that was certainly looking to primary research and also likely to, for instance, like, and attitude survey data. In a lot of the other cases, the personas that were set up, so, like, the first wave I was talking to where the category was already called out, those might be created on our customer side from, again, survey data from workshops, from raw transcripts, and so on. They’re then processing them, and we’re then building them into the system. I think with the the final piece around the respondent, the large end respondent panel stuff, that is certainly built from usage and attitude survey data primarily so that we can ensure that we’re having some representativeness across the sample in that way. And then we’re doing validation tests both on our own, but also with customers. So with the quant piece, typically, we’re comparing it to actual past outcomes they’ve done in paid response studies and so on. In the call, we’ll we’ll often, you know, with our customers, be validating that the the outputs match what they’re seeing in real respondent interviews and so on. And then we’re also introducing some capabilities to it’s I would call it validating, but a little bit different way. We we check the persona outputs against the research repository that we have on our consume customers’ behalf so that we can validate, if you will, or add to the persona chat with, hey. There’s also research been done in Market X that supports, contradicts, maybe, like, you know, just augments what the persona was saying. Look further there. Very nice. Unfortunately, I think that might be all the time that we have for questions, but we’ll we’ll send the rest over to you and see if we can get some answers.