How AI Agents will change the way brands innovate

Ryan Barry: Hi, everybody, and welcome to this episode of Inside Insights, a podcast powered by Zappi. My name is Ryan and I'm joined by my friend and colleague, Steve Phillips. Hi, Steve. 

Steve Phillips: Hello, Ryan. How are you doing? 

Ryan: I am great. I am great. We are recording this episode the day before the U.S. election, so it's still quiet with the exception of advertising everywhere. And hopefully it's a good day for the polling community tomorrow. Hopefully they're accurate. But Steve, since I've last had you on the podcast, there's been a lot of advancement in the technological ecosystem and the way brands are adopting it.

Ryan: But also in life. Um, you have a new job. You are now the chairman of our board and the chief innovation officer here at Zappi. How's that going for you? 

Steve: It's good. It's good. As you know, Ryan, I've been wanting to sort of back away from the day to day for a while and do what I'm most passionate about, which is broadly innovation, but also specifically integrating AI into our offering and maximizing the capability of us as a technology company, but also, probably more importantly, maximizing the value of our client's data on our systems.

Steve: And I think we're at this crossroads, it's really interesting spaces in the industry. Where we can take that data and really, you know, uh, 10x the value of the stuff that we collect and the understanding we have of consumers. And, and, and I want, I want to innovate in that space. I want Zappi to be at the forefront of the innovation around that space.

Steve: And it's just a really exciting place to be. 

Ryan: Yeah, it is. And so I, um, I also am doing something different. Which you probably read from some of our announcements, but I'm spending all my time now thinking about our customers. So that means two things, um, I get to do something that I enjoy all day, which is spending time with, uh, customers, insights departments, basically all of you who listen.

Ryan: The second is I'm finding that we're working together more than we used to, which is really weird.

Steve: Turns out that you need to keep innovation and customers really close. Who, who knew? Who knew? No, so it's great. It's really great. And we’re very, let's be honest, Ryan, I, and I know there's a lot of credit to you here, but we've got some amazing clients who like innovating with us and, and like, like experimenting, um, and are prepared to put the time and the effort in to, to, to see it.

Steve: And, and you know, what experiment is like, right? So sometimes it fails. And, and, and you've just got to be okay with that going in and say, yep, I know sometimes it will fail. Sometimes it will fuck up, but sometimes we'll create really exceptional value. And, um, I think we're on one of those, uh, terms at the moment.

Ryan: I, I think you're right. So shout out to, uh, our new boss, Aaron, for letting us play all day. And so he can be in the company and we can focus on what we love to do. Um, I wanted to say something before we dive in. So for everybody listening, I've been making a point the last several seasons to bring Steve in one episode a season because there's so much changing around the landscape of AI.

Ryan: But I wanted to start with something that I view as really positive and I wanted to get your thoughts on this. Ever since I got into the Consumer Insights space, there's been two constants. A passion for consumer centricity because I know it leads to growth and a frustration at our lack of ability to get out of our own way when things are moving.

Ryan: This time feels different. A couple of weeks ago, we were in New York City. Um, the truth is we had several pints of Guinness together also, which I really enjoyed, by the way. But we had a conference with a bunch of our customers, and I remember sitting in the audience thinking to myself, Holy shit, insights are actually leading a I in a lot of their work inside the company.

Ryan: Most insights people that were there are either chairing or core parts of the company's internal AI initiative, and it was very validating, I think, for a lot of stuff you and I talk about, which is the consumer data is the superpower, and AI is a great enabler, and the fact that insights departments with varying degrees of corporate governance and comfort are leading their adoption is something I've never seen this industry do.

Ryan: And so I wanted to give everybody credit for leaning in and and really driving change from the front because that's not been something I've often given you all credit for, but I think you would probably agree, right? Are you seeing the same thing? 

Steve: Yeah, completely. There was a survey from the AMA that came out the other day that said of all of the different disciplines within marketing, market research was the one that was most advanced in terms of AI usage.

Steve: And I think that's because we're in such a good position, right? So if you think about the data streams going into a CMO's tech stack. There's a lot of really big data, but that data can be quite boring and very singular. Like, you know. Do you make the button red or yellow on your website to increase clicks and, you know, yellow increases it by 0. 012 or something like that. So lots of big data that's useful. That is useful, but it's not rich data. And the thing about consumer insight data is really rich. You really understand people. We understand what they like, what they don't like, what they are, why they're attracted to brands, why they buy into a category.

Steve: The problem we've had is that individual projects are too small, right? They're 200 people, 400 people, right? So they're small data, small, rich data. What we're now in a situation is that we've got a platform where you're storing all of that data together. And if you combine all of that data, suddenly you have rich big data.

Steve: And that is what really powers AI. That's what, what's what excites the algorithms. That's what makes AI really understand a category, understand a target audience and be able to become really interested in innovating and creating new things. So we own that as insights were at the forefront of what AI models can do if we integrate that data properly.

Steve: And so it's just a terribly exciting time to be in insights. 

Ryan: It really is. And we talk a lot as Zappi around the need to be kind of systemic in your thinking, right? You answer a certain amount of questions. What's the way in which you use various technologies, vendors, and thought partners to capture that data so that the data goes to work for you.

Ryan: And, um, we obviously have the benefit of working with some amazing companies and the companies who have listened to us and are doing that are in a position now where they're actually step changing the way their company does innovation. We have several examples where you know, there's implementers.

Ryan: They're strategic consulting agencies. There's big brands. There's CMOs and CIOs who are leaning on insights to fundamentally change the way companies do breakthrough innovation. I talked about this last week in my newsletter, but Steve, I think you've got a really important vision of what innovation should look like with consumer data at the center, with AI enabling productivity.

Ryan: So take us through some of the work you've been doing with some of our more progressive customers around actually fundamentally changing the way this legacy mindset of linearity to bring the right products to market faster. 

Steve: Yeah, sure. And I should say, um, you say it's my vision. Really it's the vision of a group of people.

Steve: It was kicked off originally by a client or potential client. Um, and we've been iterating over the last year with this. So it's definitely not mine. It's a collective vision. And then, the initial idea, um, again, I said came from a client, was I've got all this data sitting in Zappi. Wonderful. I know what people like about products, what people don't like about a product, what attracts them, what type of products they own, it's all tagged.

Steve: So I've got this wonderful data. Can I go to that data and say, let's create a new product based on all the insights that I've already got, right? So I've got all this insight in this category. So we thought, oh, that's interesting. Let's try and do that. So we probably had our first workshop around this maybe a year ago, and it was basically putting the client's data in between An AI algorithm and an AI UI.

Steve: And it was really interesting. We got some interesting results. Um, so some of them were brilliant. Some of them were utterly stupid, um, and crazy. And so we've been refining the system ever since. And what we found is that if you, let's say you want to create a new beer brand for young men in Texas. If you have, So if you've tested beer brands with lots of young men in Texas, then you have some great data on which an AI can create a new product.

Steve: But that new product may be completely wrong for your brand. It may be against some of your guidelines. It may look and feel different to what you're used to. So there's a whole range of reasons why that might be the wrong idea. But it will create ideas that have consumer centricity at the heart.

Steve: And so that was the first step in our new innovation system that we launched last week. Um, what we realized is quite quickly, you can't just use a consumer insight data that we've got, you need other data and it might be social media data, but it might also be the client's tone of voice. So we realized early on, you need to feed tone of voice into the agent, uh, into the system, otherwise you won't create, um, a good, good innovation.

Steve: And so there are a whole slew of other things that we were doing that we realized, oh, if we add that, that'll help the innovation. If we add that, that'll help the other innovation. And then after a while, we realized that the more you throw at an LLM, actually it starts biasing towards the most recent stuff.

Steve: It ignores some of it. It doesn't quite work. So you have to be very careful about how you prompt it and how you manage the system, but you can create genuinely breakthrough ideas using our data with another level. So that's where we were doing a series of workshops with different clients, and we were getting really good ideas out of it, but there were frustrations in the process.

Steve: And then in February this year, OpenAI released their agents capability, and that, that has really transformed the process. Now other AI models have also come out with their own agents infrastructure, and so we're using different agents. Different situations. But basically what you can do is you take each of those agents and you give each one a job.

Steve: So let's say you're doing an innovation workshop in your company right now, you would have 5, 10 people being in that room. One person might know the category data really well, one person might know social media data, one person may know what a great concept looks like, et cetera, so they'll have different roles.

Steve: And what we're doing is creating each of those roles with agents. So each, uh, each AI agent has a specific role, and it comes together within this universe that we've created to basically mimic the process of a marketing team getting together, but doing it in, um, Half an hour instead of five days. And do it with one person from insights in control of it and managing the system and making sure they play nicely together.

Steve: And it's genuinely changing the way companies innovate. I mean, it really means you can very, very rapidly come up with lots of really good ideas with consumer insights at the heart of them. And then the other bit to the other bit to the new innovation system that we've launched is not only can we now create ideas, but we can also optimize them.

Steve: So if you test an idea in our system, you can get a market research report back, which is what you would normally get, or you can get the AI to update the concept based on that insight from the, from the test. So if you think about the role of consumer insights, always being sort of, you know, to test and validate ideas, optimize ideas.

Steve: And now we can create, test, optimize and validate ideas. So it's a really interesting use of our data. I think it makes it significantly more powerful. We can integrate other data streams, whether it's social media or, you know, connections with knowledge management platforms. So we can bring the best thinking into the innovation process.

Steve: And we're now, you know, trialing it on advertising as well. So, um, yeah, it's really, it's just a really interesting step change that we can take. 

Ryan: Yeah, it is, right? Like, I mean, harnessing what a company knows and then being able to recreate cross functional dialogue without cost, time, scheduling, but bringing in, okay, we know that we can make this.

Ryan: It's feasible. We know that it resonates with consumers. Great. And we know that it ladders to our principles as a brand really quickly. I mean, we did this exercise for fun last week, but something stuck out to me. There's a woman on our team named Pooja. I love Pooja. She's one of my favorite people. And she created a brand for Halloween.

Ryan: And what I loved about Pooja's brand is It was Pooja. It had her design aesthetics. It had her tone of voice. So if Pooja was a company, Pooja is somebody who I know very well, but I was like, that's Pooja's company. I know that that's relevant for her, but she knew it would resonate with customers. And so if all of a sudden you're jumping from a bunch of random ideas, most of which you have already tested.

Ryan: I mean, I've said this to you for years. Ideas are, are free. It's execution that matters. But if we can leverage what we know to say, here's six that have legs now tell us how to optimize them. Well, then when we go talk to customers, we're asking things we don't already know the answer to first and foremost.

Ryan: But I think that the flipping that Steve's and talking about is exciting for me because from idea execution, we know, can we do it? Is it resonant? Is it going to sell and is it in line with our brand values? Now we can get to the fun part, which is actually trying to bring this thing to market instead of doing a bunch of your market research projects over the course of you know, 25 weeks because we know the hardest part of bringing products to market.

Ryan: We know this from our own work at Zappi, but also we consult with big companies all over the world is okay. You got a good idea. Now, how do you execute it? And so I'd rather see companies spend more time in the execution stage than the market research stage gate process. And I think this agent concept is starting to really connect that, but it all works because you have consumer data at the center.

Ryan: The agent's useful cause it's going, I know your supply chain governance. I know your brand guidelines. And I know that young men in Texas like beer. Um, now, you know, I have a son who's nine years old. He's quite a bright young kid, but he is nine years old. And, um, we created a baseball club brand the other night while watching the world series and we tested it against other sports apparel brands.

Ryan: And by the way, they're for sale at Dick's Sporting Goods. No, I'm just kidding you, but if you want to ask my son, you know, like he would love to take some capital from you. Um, but I'm starting to see this there. And I think this is a trumping of a lot of great innovation in this industry. It used to just be, we need idea screening faster.

Ryan: We need to test concepts faster. And I mean, it's one of the biggest commodities in this industry now is overnight screening. And so we wanted to reinvent the whole thing and say, well, screw this. Let's flip this. Let's let the agents do the grunt work. Let's get the people in the room when we have something we know is going to work.

Ryan: Uh, so, all right, so what's next? You're starting to drive this with a couple of really big brands. We obviously just to the world this last week, where are you taking it next? 

Steve: Well, the interesting thing is it's again, the, the, the push that the clients are making, so that they're looking at this and they're saying, okay, so I can come up with a new product concept that I know has consumer thinking at its heart, right?

Steve: So it's designed with great consumer thinking, but I can come up with a new product idea in 15-20 minutes, not in five days. Well, then you start saying, okay, how do I, if that's true, how do I change my innovation system? So instead of, you know, big workshops coming up with ideas in five days and testing them and then coming back to them a month later, what if I create 200 ideas in two days?

Steve: What does that world look like? And you start thinking, okay, maybe you can, maybe now you can create 200 new concept ideas, all of which are based on consumer insight. Maybe one of our clients said to us, Okay, what I want to do is create 200 in the first day, use synthetic respondents to whittle down to 100 or 75, then I want to test 75, I want to get rid of half of those, and then I want to optimize the remaining 30, utilizing the AI.

Steve: So on day three of my week, I've created 200 ideas, I've got down to 30 that are really strong, and I've optimized almost 30. And I've done that with a team of two people. Well, that fundamentally changes your innovation pipeline process. I mean, it's a proper behavior change that you have to start thinking, okay, what does that mean to the organization?

Steve: I can innovate this quickly. Um, the ideas are strong, maybe I can double the number of new innovations I bring to market. Maybe they're You know, 50 percent, 75 percent more successful so I can, I can put more of my dollars into innovation because I know that the ROI is going to be much stronger than that.

Steve: Maybe I can pull some away from the media because I'm making enough of the innovation. So it just changes the way you think about your ability to innovate and the power of that innovation. So it's just, it's just really exciting. I mean, we're still at early days. Doing this for a year, but the process of bringing that to market from the client side just takes time.

Steve: But it's just, it's really, it's really interesting. They're really stretching our thinking. 

Ryan: Yeah, I think it's exciting. I mean, everybody's been rolling out the old Nielsen, 50 percent of products fail chart from 1987 for a long time and faster, cheaper mousetraps didn't change the problem. So getting our thinking embedded, leveraging what we know, creating a bunch of ideas, using data and consumer feedback to optimize them, to get the components of the product right gives us a better chance at execution than we've had. And it does leverage a lot of the thinking that you get in for the lean startups, which I know a lot of big businesses have been trying to integrate, but they haven't been able to. But in three days, you can have concepts, you know, work that are optimized with consumer feedback, and you can spend your time bringing your stuff to market.

Ryan: But as ever, the people and process change around that is paramount. But this is where I think for the first time in a while, Insights is cool. Insights are driving this change. And insights is creating pull at the C suite because they're bringing the consumer data in an on demand way, not an agile way, in three clicks, you're getting answers back.

Ryan: And so to Nick and Dee and some of the other mavericks who are driving this, um, they're getting a lot of credit internally, but I just wanted to shout you all out because, um, you're pioneering the future. Um, if you want to learn more about how to step change your innovation process, we'd love to help you.

Ryan: Reach out to steve@zappistore.com. As you know, we both are no longer responsible for the spreadsheets of Zappi. We are responsible for the customers and the innovation. So he'd love to talk to you. I'd also love to invite all of our listeners to join our connected insights community. It's a safe space for you to share ideas.

Ryan: We have some exclusive content coming down the pipe. We have Richard Shotton, who's going to be doing a really cool chat around how to use behavioral economics inside your business to drive change. We have Lenny Murphy, who's a good friend of both Steve and I, on to talk about industry trends and how they relate to corporate researchers. There's a lot of good stuff coming. If you want to get in that game with us,  you can also reach out to me at ryan@zappistore.com, we'd love to have you. Ee'll be back for more fun updates. As you just heard, Steve's now playing with synthetic data and advertising. So more to come, my friend, thanks for joining.

Ryan: It's always fun to have you on the show, Steve. 

Steve: Thank you, Ryan. 

Ryan: Bye everybody. Thanks for listening.