Webinar: Now, Next, Future CPG; how Gen AI will transform the industry

We'll get started since we know everyone's time is really precious, particularly if you work in cpg. You probably have a really busy day ahead. But we're really grateful for everyone joining today. I'm Jo McKinney, Chief Growth Officer at BOI and I'm really happy to be here with my colleague Natasha to talk about the work we've been doing in CPG and the future we see as AI has transformed so many aspects of the industry. We already and will continue to do so. If you don't know us, we're Board of Innovation BOI and we're a consulting and technology company headquartered out of Belgium. We're dialing in from New York and we're reinventing growth for companies with autonomous innovation and transformation. We truly believe that our work is becoming much more autonomous and all of us as leaders are going to be more and more responsible for AI driven automation across a lot of workflows. And what does this mean specifically for CPG brands and leaders and those of you gathered today in the audience? Well, high stakes, high complexity, massive scale mean AI really can significantly impact both the front end and the back end aspects of the businesses that we're all working in. So today we're going to cover a few things. We're going to talk about what's happening now, what's happening next and where we see the future of AI and cpg and we're going to kind of journey into some future worlds that we really look forward to getting your thoughts around. So let's start with just a 2 seconds of CPG level setting. Perhaps for those of you who are very familiar or not as familiar, but I assume if you're here today it's because you work in cpg, so you'll probably get it. But CPG really is in a league of its own with billions of daily transactions which really rapid innovation cycles, global supply chain challenges, you name it, there's an erosion of brand loyalty, it's a really hyper competitive environment and it makes it really ripe for AI driven transformation. It's really hard to think of a more complex business model than one you find in cpg. And because Gen AI really, if you can go to the next slide, Natasha. Because Genia really excels in the most complex spaces, it's really no wonder that some of the earliest use cases for Genai have been in cpg. We've seen its impact everywhere, from backend applications and things like customer service and supply chain optimization to front end innovation, marketing insights and new product development. In fact, go to the next slide. Natasha, all the big players are in the mix. These are just a few recent headlines, but everyone's doing something really interesting in this space. And As a company, BoI has really been on the forefront of leading in this space. So we're really excited to share what we've been up to as well. And we're going to start really close in talking about innovation, which is one of our favorite areas. I'm going to pass it over to Natasha.

Awesome. Thank you so much. So as Jo mentioned, there are so many use cases of generative AI within cpg from the front to the back end. And as she mentioned, really the example of what is happening today in cpg, we're going to focus that conversation with within innovation. And so the work that we've been doing with CPG companies within innovation is to really make front end innovation more efficient and effective. And we're doing this through kind of innovation engines and we'll go into what that actually is. But really these innovation engines completely change the way we invent products, services and experiences. And we've done this by leveraging our years of innovation roots and traditions at boi. And we've really kind of transformed a human only process into a human and machine process with these innovation engines that basically automates everything from insights generation, concept generation, validation, product design and development, and of course scaling and launching. And I'll bring this kind of innovation to life in a demo very, very shortly so that you can actually visualize and understand what we're talking about. But I just want to quickly highlight here that engines go beyond just innovation. What Gen AI allows us to do is bring forward some of the feasibility, the viability questions and considerations that sometimes either get overlooked or we don't have enough resources to explore holistically in human only processes. So it really impacts the entire business very deeply and simultaneously. So why should you care about kind of innovation engines and what's happening now with innovation engines and how does it actually impact the world of innovation and cpg? Right. So these innovation engines enable you to move through the entire innovation process within one to four weeks using both kind of human and AI creativity and ingenuity instead of a process that takes kind of three to six months. And this obviously has so much value. So I'm not going to like target kind of the main value areas, but the one thing that excites me about it is that you just have more time to to do things a bit more strategically, creatively. You can explore more consumers. For example, we traditionally in Innovation only target or kind of validate concepts with Gen Pop. But now that you have more time and resources and money, you might want to kind of chat to a new subset of consumer to really enhance that concept that you're building. These engines also allow you to explore kind of multi market trends, social listening category, competitive analysis, almost autonomously and instantaneously. So previously doing things like a competitive research or social listening sometimes is targeted to one market. And that in and of itself is really time consuming. But really you can do this instantaneously and you can do it globally. Right? You can generate hundreds of concepts and test them to surface the best ones that meet the needs of your consumers instead of forcing a prioritization in early stages. So all to say is that these engines have a lot of benefits across innovation and beyond. And if you were wondering, these engines do work. We built a couple of these engines that are kind of CPG clients. But an example that I'd like to share with you is an innovation engine we built at a global CPG company. And we wanted to test out, does this thing actually work? Does it produce kind of quality results? And so we did a MaxDiff experiment where we took 10 engine generated concepts and then 10 concepts generated from traditional kind of human innovation processes. And these concepts generated by the human was previously tested with the Nielsen Spaces concept test and they all were rated superstars. So they were really great kind of human concepts. And we wanted to pit these kind of 20 concepts against each other. The engine generated and the human generated. And what we found in this maxted test is that 4 out of the 10 concepts generated by these innovation engines were in the top 10. Really kind of proving out that what you can do in just a couple of weeks with AI and human provides the same, if not better results than human, only human methodologies. So without further ado, let's kind of really dive into what do we actually mean by an innovation engine and what are we currently doing now that evolves the CPG industry with these engines? So I'm going to take you through a demo that we put together. This of course, is in a kind of real engine that we built. We just wanted to show you what kind of this looks like for illustrative purposes. So this engine has three elements to it. It has the insights, the concept and the validation engine. And today we'll go through each part of them in kind of a more linear format. So if you want to kind of create insights, the first thing you want to understand is what data do you want to include in your mix, you can include your proprietary data, for example, you can include scientific data, social listening data, your consumer research data, and upload any documents that you would like to. Right. So we can kind of click a couple that we to want to work with. We then want to understand. All right, what do we actually want to do? Do we want to develop opportunity spaces or Personas or detect trends? For now, we'll go with opportunity spaces. And then the engine basically allows you to kind of navigate it a little bit more by kind of adding in prompts to say, focus on this area or this market or whatever kind of you want the system to navigate towards. You can kind of provide that. And as it's doing that and as it's generating these opportunities, I wanted to call out that generative AI models are very good at clustering data. So think about your chatgpts, your clauds of the world. They're really good at taking a bunch of data and clustering them. The unlock in these engines is how do you make this data clustering into something that is insightful? Basically going from data clustering to insights. And what we've been able to do is really codify human intuition and really stretch kind of data to kind of beyond what the clusters actually mean into something that's a bit more insightful. So here in your insights engine, you have various kind of insights. Like I mentioned, these are super generic kind of content for illustrative purposes. But you have kind of your opportunity spaces that you can, that you can kind of review. You could also kind of take a look at the Personas if you kind of have chosen to generate various Personas. So that's kind of your insights engine, right? Then you go into your concept building within concept building. You want to kind of do your setup. So what division are you kind of part of? This example is a food and Bev Co. Example. And so we pick a potential division that we want to focus in on. We select a Persona that we want to kind of generate concepts for. And of course we kind of pull in the opportunity spaces that we've just generated and focus in on one of those opportunity spaces. By clicking this, you basically take all the insights that you've generated into this concept generation. You go ahead and you create these concepts and you do some ideation. The call out here is similar to what I mentioned before. AI is great at clustering data. So the goal here is, how do we go from just kind of average ideas that potentially exist today to something that exists beyond the category or something that's really kind of novel or disruptive. And the way that we've been doing it is by using agents. And spoiler alert, Joe's going to dive into what agents actually are. But essentially in this kind of ideation use case, excuse me, we basically create agents that represent chefs or ingredients or R and D. And they kind of work together to really stretch these ideas into something that's a bit more novel. And so you can kind of generate these ideas which will generate an image for you. You can kind of edit them by yourself. You can ask AI to generate them if you'd like. This is a very kind of simple interface of what this could look like. Then of course, a concept is nothing if not validated. And so there's various kind of ways in which validation occurs, as I'm sure you all know, all the way from synthetic to actual consumer driven validation. We don't believe that you should only be doing kind of synthetic like testing. We really do think humans play a huge key role in this. And so you can see that we do have some validation techniques such as AI moderated validation or bases testing and things like that, that would require you to kind of connect these concepts that you've just developed and go to the other tool that provides these validation experiment techniques. But the one thing that we do recommend is because generative AI allows us to generate lots of concepts, we recommend doing synthetic validation first before spending money on kind of consumer focused tools. So for those that don't know, synthetic validation basically refers to the generation of artificial responses from artificial, artificially generated kind of users. So for example, if you want to test your concepts with a subset of a Persona or even a, or even the genpop, you can basically generate synthetic consumers that reflect the general population. You can test your concepts with them. I think it was last year we ran a project with another CPG firm where we basically had a bunch of concepts and we wanted to see if we test them synthetically and if we test them with humans, what is the delta or the difference between the responses between synthetic and consumer and directionally synthetic feedback gave us really good results and the same results as human beings. So this is a really good and effective and efficient way, way to do initial validation. And so you can actually do synthetic validation within the innovation engine tools. You can kind of run the test, choose to do a desirability test, for example, and then you will essentially come out at a response that basically tells you what consumers like, what they don't like. It'll give you a couple of scores and how to kind of improve the entire company concept. So that is the engine. There is a lot more I can dive into the engine, but of course we don't have time to do that today. Happy to kind of go into it in more detail if you'd like. Feel free to contact, contact us. But I think just to kind of wrap this section up of what is happening now in cpg, I just want to highlight a couple of things. The first is innovation engines are clearly reinventing the way we innovate. And not only that, it's really transforming the human role from being like the people that actually execute to being someone that is a bit more strategic and more creative. These engines are also built so that humans are in the loop. And what do I mean by human in the loop? I mean that humans are the one to prompt, edit and approve the outputs that the engine generates, which is really important, especially as we're early in this technology. And that also means that the human role transforms from being transforms into being something that's more editorial and more critical. Lastly, humans in the loop systems also mean that humans will activate the insights, concepts and validation generation engines like I've just done. But in the future you can also imagine that, and it's very near future actually that you can imagine the engines actually proactively nudging the human being to say, hey, there's an insight that or a trend that's happening in this part of the world. We actually think it's relevant for your market here. Or there's this new concept that I, that I've developed and I think it's going to be really competitive in your market. So this really just means that we're able to be a lot more efficient in the now.

Great. Thanks, Natasha. I think that was a really great demonstration of how with innovation, AI is really delivering a lot of efficiencies in existing operating models and with existing products and services, which is the lower quadrant of this evolutionary diagram you're seeing here. But we see really a massive evolution in autonomous work streams on the horizon where more and more we're going to be pushing towards the top part of this, this quadrant where we're getting more AI native operating models and AI native products and services and delivering really autonomous breakthroughs. So while that engine that we've built and that we've been working on is really providing great value, it's really living in this lower part of the quadrant right now. But we're going to now talk about what's next and how we're pushing towards these more Autonomous work streams and what we call an agentic future. So is your first question when I said that, like, what is an agentic future? If, if it was, you're not alone and like, let's talk about it. So first off, and I'm just going to do a little level setting on language because I think it's important given how fast all this technology is evolving and how everyone has sort of a different lexicon for talking about what they're doing. And I can see in the chat there was also a question around how is this different from a custom GPT? And so we're going to talk about that really quickly. First off, what is an agent? An agent is an autonomous software entity, so it processes information using algorithms, it takes actions against specific goals, it operates, as we said, with a level of autonomy, and it makes decisions and learns from its interactions to optimize outcomes over time. So it's not just a custom GPT, it's not just a chat. It actually functions autonomously and makes actions. So when you think about an agent, an agentic future assumes that AI agents are actually working together autonomously, making decisions and shaping outcomes in what we call self regulating systems. So this is sort of the world now that we're going to start living in for the rest of this presentation. And while many of us have been building custom GPTs, as mentioned in the chat, in our daily work, those are definitely foundational elements of agents, but they're really just building blocks because to really be an agent, it needs to work with more autonomy, it needs to learn and act without direct human input. And right now a lot of what is happening in custom GPTs involves a good deal of human interaction, as Natasha showcased and mentioned. So when you think about that and how the agent landscape is currently evolving and pretty quickly, we know right now agents are really operating within these defined parameters and human supervision, as we said. And for the most part, most of us are using gen AI tools alongside traditional ways of working. So, you know, we're incorporating genes into the ways we've worked in the past rather than them operating in an autonomous system. So very soon though, we think AI agents are going to operate with autonomy and proactively and we'll be making decisions and taking actions without constant human guidance. And that's what we're going to now dive into a little bit deeper and again, feel free to leave questions in the chat. Sorry, we'll, we'll try to get to them towards the end of the conversation. So in the near future we're sort of Imagining there would be an engine room of data tools and agents that are supporting all the teams in our workspaces and that they'd be working simultaneously and autonomously to perform tasks, make decision and interact with their environment, just like humans do in a work environment. So imagine on the left hand side here is the knowledge base. So any, any types of data that the company might have internally or externally, brand assets, frameworks, all your strategic underpinnings of your, of your business form the database. And then within the autonomous system, there's various agents that are performing tasks and making decisions and interacting with their environment and with the humans that are also in that environment. And of course leveraging a lot of different forms of technology to do that. And the use cases really are across the boards in so many different aspects of the business. And we're going to delve into a few of those right now. So when we think about where agents could impact cpg, we think mostly about wherever there is complexity. The most complex place is the best place for AI. So let's look at a few parts of the business where agents could step into all kinds of roles. Whether it's making marketing more personal, let's say, or streamlining your supply chain, or helping companies become more sustainable. We're going to look at how agents could thrive and actually drive in processes that affect areas of desirability, feasibility, viability and sustainability across the business environment. So starting with desirability, which is, as most of us know who work in cpg, it's all about making every customer interaction more meaningful, more insightful, more customized, more tailored. And this has really been the focus of a lot of the early CPG work to date with AI, but in a more agentic system, with more autonomy. Imagine what it would mean to be like always on with some of these things, whether that's insights, marketing, copy development, building products within demand spaces. Imagine that you have agents that have been trained based on all of the expertise of humans and the data that exists within your company to do these type of tasks all the time and without a lot of interaction and supervision. If you move on to feasibility, which is really about operational capabilities and resource management across the business, a supply chain optimizing agent could dynamically adjust operations in response to real time data changes in demand, disruptions in logistics. A technology scouting agent could identify emerging technologies that could be integrated into the business process, making sure your operations are always ahead of the curve. And we've already built an ingredient innovator. Natasha was showing you one of our innovation engines and an Ingredient innovator could actively recommend new effective and ingredients to either enhance benefits or address costs of goods. So there's a lot of different work that could be done in feasibility in the background in an agentic system. Viability, which is up next, involves ensuring strategies are financially sound and market ready and can cover everything from pricing to global market dynamics. You know, you could simultaneously have an agent working to reduce production costs, while another could be focused on generating ideas that would allow for premium pricing, both of which are really important to the business model and then sustainability. Obviously this is about environmental impact, but it's really about more than that. It's about creating business practices that can endure and evolve. One of the more common discussions we've been having lately with clients is thinking about an agentix system in the circular business model, planning and executing workflows to reduce, reuse and recycle from the onset and really building that in via agents. So, you know, when you think about this, I think the real magic happens when the agents start working together. So if you could all just imagine in your current workplace a system where everything's connected, your marketing, supply chain, pricing, sustainability efforts are all feeding into each other and where an autonomous process happening allows the humans in your organization more time for strategy and creativity. I think the use cases are easy to think of when you consider how complex the CPG industry is. And I'd love to see in the chat the use cases that pop to mind in your own business. You know, what are the most complex aspects of your business that are difficult to solve for? And what would your ideal agentic future in your business include? If you think about all the things that could be happening autonomously and in the background, I mean, this is. We spend a lot of time thinking about this, where are rebuilding a lot of these types of agents. And because of that, we've been giving a lot of thought to like, well, what could the world look like in the future? And I'm going to hand it back over to Natasha to talk a little bit about that.

Awesome. Thank you so much, Joe. There we are. Right, so Joe kind of provided a glimpse of what the syntax future could be. The syntic future, coupled with kind of the evolving technologies within AI, generative, AI, robotics, etc. Kind of all lead us to a world where things we deemed impossible before, or potentially even a little bit sci fi, are now more kind of possible to achieve. So we want to imagine what are the previously impossible challenges and complexities across the CPG ecosystem that we can potentially kind of solve for now. And we want to kind of do that through this element of what does the future, what could the future really look like for CPG? And as I'm sure kind of, I think there's 126 people on this call, I'm sure most of you are experienced CPG folks or have some affiliation to the, to the industry. And if I were to ask you to imagine what that future CPG agentic world could look like, I'm sure you'd come up with hundreds of interesting potential futures. Again, don't have time to dive into all kind of different angles, but kind of we've picked a couple of potential futures across the consumer, retail, marketing, media, innovation that we can really dive into. And as we're going through these potential futures, what I'd like for you to think about and enter into the chat is how you think these potential futures might impact you as a consumer, as an employee, and within your company as well. And so Jo, I give you absolute permission to kind of interrupt any future that I'm talking about with a kind of interesting comment from, from one of the folks on this call. So let's start off with the consumer kind of. I have two kind of consumer potential futures. And in the first potential future, imagine a world where most consumers, obviously not every kind of consumer is going to have this because of various inequalities and access and everything. But if most consumers had access to their own personal AI agent and that AI agent was able to anticipate human needs using predictive algorithms, these agents can draw on real time data such as weather, traffic, location, your kind of calendar, your health metrics, even your emotions to really provide you with recommendations that go beyond hyper personalization and in some instances also go beyond recommendations and into autonomous decision making. So if I were to think of myself as a consumer, for example, I would love my grocery shopping list to be automated and purchased for me, especially if something that I'm something that's not in, in stock right now, I'd love for that to be kind of purchased autonomously for me. But other decisions, for example, I wouldn't want that to be automated. I'd rather have a partner in crime to kind of discuss and navigate that decision through. But what does this really mean? What does going beyond hyper personalization and autonomous decisions actually mean for a consumer? The way that we're kind of potentially seeing this future kind of play out is that the shopping experience becomes a more passive one than an active one, which is the world in which we live in today. The next potential future Also kind of related to the consumer is a world where consumers might wield unprecedented control and power over their AI agent. And what I mean by that is I can kind of program my agent to do certain things that I would like it to do. And by doing that, that kind of personal agent of mine acts as my personalized gatekeeper, filtering out kind of unwanted ads or any unwanted content online. So for example, if I am going continuing with that grocery shopping example, I want to only see healthy foods, for example, or foods that are ethically sourced. I might kind of program that into my agent. I might only get recommendations for products that fit that criteria. And so what we're really seeing is that the balance of power could tilt from kind of algorithms, from kind of companies towards the consumers who actually now dictate not just what they want to see, but how they want to and when they want to interact with brands. And so what we're kind of potentially seeing as a challenge for CPGs in the future is that CPG firms will have to break through these kind of AI driven filters by aligning more closely with kind of consumer values and preferences and behaviors. And we're seeing kind of signals of this happening today. Right. Our algorithms on social media are kind of filtering out things for us. But imagine if us as consumers had control of that algorithm. And I think that's where this future potentially comes in. Cool. The next potential feature is within retail. So there could be this shift from traditional screens and keyboards to more seamless interactions. Now, I'm not talking about kind of AR VR headsets, but rather what I'm talking about is voice command, gestures, biometric signals, kind of more seamless interactions in, in those kind of formats or channels that basically make shopping a lot more seamless and a lot more effortless. As effortless as a spoken word. And again, we're seeing a lot of it happening today. With Amazon's Alexa, for example, we can just command Alexa to do something and that technology will do it for us. And so we can see this growing and exponentially growing with kind of generative AI. And so that basically means that the lines between browsing, marketing, buying, kind of all blur and shopping becomes a lot more continuous. And so when we're thinking about this potential future, we're really thinking about how do brands in this future keep up with consumers who are always connected, they're always engaged, and really whose every desire could potentially be very easily fulfilled. The next potential feature is within marketing. So in this kind of future, we're making the assumption that every brand has their own AI agent. And we're kind of saying that these AI agents are actually built on the same or very similar LLM models. So again, think of like OpenAI and philanthropic llama, et cetera. It's all built on the same underlying model, which basically means that we're at risk of a future where everything is homogenized. So messaging, communication, consumer interactions all kind of become the same or very, very similar. And so the question we're asking ourselves is how do we break free from the sameness? And I'm 100% sure people on this call have experienced this when using things like chatgpt or Clawed. And I can say that with confidence because we experience the same issues as well of just this kind of sameness that occurs. And there's ways in the future that this can be overcome either through technology, but the way in which we potentially seeing this, we see this is maybe our competitive edge is actually in our people and in the people that we hire. People who are extremely creative, they're out of box thinkers, they're risk takers, they're people who are able to disrupt the uniformity that these algorithmic worlds bring to us and introduce something that is actually truly human and simple, which is uniqueness. So maybe in this future we actually go from AI and tech to actually a bit more of like human, human kind of responsibility and creativity. Our next potential future is within medium. So in this kind of potential future, what we could be seeing is centralized social media platforms like the XS of the world, Instagram, TikTok, et cetera, could all crumble under the pressure from decentralized AI powered ecosystems. Now these ecosystems and these new systems, the centralized, sorry, the decentralized ones, could allow users to actually control their own data and monetize their engagements, shifting the power dynamic from traditional platforms. And in this kind of fragmented media environment, content is also created by users and AI and it's kind of seamlessly woven into virtual communities, immersive experiences, new kind of channels like the ones that we previously mentioned, one that is voice command focused for example. And really audience kind of may detect what they see with personal kind of AI filters offering hyper personalized permissions, mission based media. And again, this is not far off from where we are today. What's happening today and the signals we're seeing today is really kind of going to be supercharged by gen AI, by this agentic future that's going to lead us into kind of a more fragmented media environment. And last but definitely not least, the innovation engine that we just kind of showed to you earlier on. And these kind of agentic futures that Joe went through, that encompassing encompasses everything from desirability, viability, feasibility, sustainability, These all enable us to flood the market with endless new products based on the newest trend, the newest consumer need, the newest kind of competitive opportunity. Right. What that basically means is that what that potentially could actually lead to is that creation could outpace consumption. And there will definitely be brands that kind of lean into this trend where they create more than people can actually consume. The question you need to ask yourself and that we're asking our clients, is, do you want your brand to kind of race to keep up with this, or do you pivot to kind of generating or creating fewer, more meaningful products that truly resonate? And that begs another question, which is just because you can do something, should you do it? And with that, Jo, I'll leave it up to you to bring us home.

Thanks, Natasha. So the next real question is, where do humans fit? And if you haven't been started in this area, where do I start? And the more we lean into this future agentic world, Natasha literally just showed you a handful of imagined scenarios, but we've been playing that out across every aspect of the business. The more we think about how we need to rise to the occasion as leaders and as humans. In a meeting we were in last week where we were imagining this type of agentic future, the client who was a leader in AI and CPG said, better AI needs better humans. And what he meant, I think, is that as this technology advances and we contemplate how to use it and leverage it, we really need to be the best versions of ourselves as humans and as business leaders. And I really couldn't agree with this sentiment more. Our role is going to evolve from executing tasks to setting strategy and driving transformation in a world that's increasingly influenced by AI. And we really need to think about that long and hard. And as Natasha just mentioned, just because we can doesn't mean we should. AI really has the power to revolutionize our industry, but we have to balance that innovation with ethics. And we can harness AI's potential really responsibly. Protecting consumer privacy, for instance, avoiding biased outcomes, ensuring transparency, particularly with our consumers and customers, and protecting against cyberspace threats and data threats, and working to reduce our environmental impact. These are all kind of non negotiables. If we want to grow ethically and if we get the balance right, I think AI can drive really incredible growth. But if not, we can really undermine everything we're trying to build and what we hold as important both as humans and as leaders. It's really, I think, just an important inflection point. So we would say as we move forward, let's move forward really thoughtfully. You can go to the next slide. Natasha. We think against sort of three critical areas when we think about how do you get started or how do you move in the direction of an agentic future. Obviously you need to set a strategy envisioning how to act in an increasingly autonomous world. What will that mean in your business? What will it mean in your group and your department, with your staff, all of that. You need to build products that are, and engines that are working for you always on, across aspects of the business and innovation growth, whatever that might be. And then you really need to drive the transformation. You can't just use new products and services or new tools. You need to transform your human AI operating model to really maximize the potential around all that. So the starting point as we see it, a few things to consider. First of all, and I think I've said this three times already, and it's quoting one of our colleagues, Jeff, you know, start where things are most complex because that's the place where AI thrives and humans need the most help. And I think that's why we've all probably gravitated to certain chat based AI tools that we're using to sort of take some of the complexity out of the daily parts of our work. But I really think we do need to start thinking beyond what today's tools can do and plan around that future of where we think the technology is going. And as Natasha just shared, AI gives us, us a chance to create previously impossible solutions. And that's really the most exciting thing about this, is really dreaming about being able to do things we couldn't do before. Obviously, as we said, it's not enough to have great tools. We need new ways of operating and new ways of working. Otherwise we're missing the greater opportunity across the boards. So as a, As a company, BoI is spearheading a lot of groundbreaking work. If you're interested in what Natasha showed you earlier in terms of an innovation engine, we could build that, that for you today. You could have that up and running within a few weeks. But we're also doing what we're calling these impossible strategy sprints, which are like four to six week processes that give you great growth strategies, innovation concepts and a reinvention blueprint that really addresses all the aspects of the business across the kind of transformation that we think needs to happen to support this future we're envisioning. So if any of that is of interest, obviously reach out if you'd like to discuss it. Our emails are available and obviously we'll be sharing this webinar. But we'd love to now go to the chat and answer some questions, have a little bit of discussion, because I know we were really stretching imaginations a little bit about what a future could look like. And we're so excited to hear from the hundreds of people who are here today around, you know, what is interesting to them, where are they currently working? Where do you see the applications for your business? So feel free to dump some stuff in the chat. Natasha and I will now also go back and look at what you've been saying over the last 45 minutes because it was hard to read and speak at the same time. Natasha, if there's any specific questions that you see in there so far that you'd like to address. I know there was some early questions as you were talking about synthetic testing early on, if you maybe want to describe that a little bit more in depth.

Yes.

Questions on how does synthetic validation work?

There's various kind of fidelities of synthetic validation and there are definitely more experts at by to kind of run. Run through this from a technical standpoint. I'm not on the kind of technical side of things, but there's kind of various ways that we can do it currently super easily. The way that you can actually think about synthetic feedback is you can literally just construct your Persona of Gen Pop through kind of a GPT chat. Very, very simply. In that case, what you're seeing is you're having a conversation between a one consumer and yourself and you're testing out a concept with that one consumer that you've synthetically created within that chat. So when we're actually talking about synthetic validation, it goes far beyond that that we're saying, let's say you want to speak to consumers who are between the ages. I'm making this completely up between like 30 and 40 years old. They're based in the States and they have ABC qualifications. Right. You would basically kind of program that in to your generative AI system and that would essentially generate kind of of hundreds of consumers with those attributes but that are different from each other and that replicate kind of the Gen Pop environment. So basically you have kind of this list of hundreds of consumers all between the ages of. I think I said 30 and 40 with ABC kind of qualifications. And then you can ask them various questions. And these questions essentially can be more quantitative. Questions like, like intent to purchase. It could also be more qualitative questions to give, to give, to get feedback. By now you can't do this using just kind of your chat, kind of your chat, like your chat GPT for example. You'll have to kind of create kind of an API to link it to an Excel. It's a whole, it's a whole spiel that I won't pretend to get into because our technical experts at, by kind of do this, but essentially you're able to then test your concepts at scale. So that's just a very kind of high level overview of how it could potentially work.

Great, thanks, Natasha. I see there's also a question around the innovation engine in terms of what if you don't have any data of your own? And I think that's a really interesting question. Obviously there's some very, very large companies that we've been working with who have advantages of years worth of in data, whether that's segmentation studies, sales data, you name it. But the interesting thing I think that we're building into these innovation engines and Natasha, feel free to correct me if I'm, if you have more to add here, is that we have the ability actually to bring in a lot of outside data as well. And that's been really interesting. I mean we did some really interesting work where we were incorporating trends in the moment based on looking at conversations in TikTok and Reddit for instance, and having that inform insights lots of different other data sources that we can purchase if they aren't things that are readily available. But Natasha, do you have anything to add for that question?

No, I think that's exactly it. I think you can just tap into kind of the wealth of data that's out there and it's existing. The one other thing that I could offer is if you were to add kind of more synthetic data. Right. So if you were to say my consumer is this kind of profile, can you generate some insights and needs based on that? So we can maybe even kind of generate some synthetic data to get you started. You would want to validate that however, in the future. So that's just a potential add on that you could consider.

Excellent. I know there are a number of questions on the technology side of this in the chat and I want to say that we have more webinars coming up covering this topic where we'll dig more into the first steps and implementation of the sygentic future, led by our Managing Director of product, Amir uk. So I would encourage you guys to bring those tech questions to that conversation because Natasha and I really, it's not our, our specific area of expertise to answer. So I'm going to focus in on some of the other questions that are a little more in our wheelhouse. Chris just asked a question about how difficult do we find convincing new clients in some of these areas before building and leveraging new approaches like synthetic users, research, validation. I think that's such a really interesting question. I will say as someone who's working on the growth side of the business, I'm spending most of my time and days in conversations with people who I really respect who have long careers in really important areas in cpg and they have proven ways of doing things. They also have a lot of historical data and benchmarks because of the way they've done things. So this is one of the most critical questions in many ways like how do you get started thinking about new ways of working and what does it mean for all your old ways of working? And I think there's not one perfect answer to that question except that I truly believe that people with the most experience need to be the ones who are most experimental right now. You know, we have great knowledge from our years of human process around what good work looks like in whatever that area is, whether that's insight work or innovation work or marketing or whatever it might be, supply chain management. So what I've really been encouraging people is that take your most expert people within your organization and let them be on the forefront of experimenting and finding ways to use this. And often what we are doing is standing up some of these more AI driven processes right alongside the traditional human process and looking at the results. Natasha showed you an example of that. We're doing that right now for a large company in the OTC space and I think that's really the way there. I don't think there's a lot of businesses that are just going to jump to whole new ways of working without having that experience. And I think we're doing a good job of getting people comfortable and giving them proof of concepts, work streams that allow them to really get their feet wet, get answers quickly and measure success against the metrics that they have historically found really important in their business business. So I hope that answers that question. I see, yeah. People saying this is 50% of my day overcoming convincing people. Yeah, I think, you know, there's, there's a lot of people, I think who are, who are excited and nervous at the same time. That's sort of where we are and rarely have we seen Such a fast level of adoption and transformation in industry. So, you know, it's really understandable. Natasha, I don't know if you have anything to add to that.

No, I completely agree. I just think we found success in starting small and just like Joe mentioned saying, hey, here is the metrics for this small little task, right. It could be just using AI in your daily, in your day to day and showing what it can do and then progressing it from there so you don't have to go from 0 to 100 to build an autonomous innovation engine. You kind of demonstrate what it looks like. You build a no code solution, you build like an mvp so you're slowly, slowly kind of investing and showing the value across the organization. But again, we love having these conversations. So happy to also do this one on one if you would like to reach out to us after this.

Yes. As a whole company, we just recently did an offsite where we did no code building across the entire team developing early agents in areas to solve our own business problems. So, you know, we are both, I would say, leading conversations with clients and experimenting on our own as well to make sure that we're, we're in the same mindset that the clients that we're working with. I see a question in the Q and A about wouldn't fine tuning be a functional solution to avoid the homogenization of future AI solutions? I think the, the point that Natasha was making about the future homogenization is that if everyone is optimizing their algorithm, ultimately that sense of fine tuning could take us to an area where everything looks the same. And I think we're seeing that already. I, I know I personally am as a career marketer seeing so much social content, for instance, that has this very similar look and feel as so many people have gone to ChatGPT to write, write, post, copy, or to design things like that, that things are starting to look really the same. So I'm not sure the answer is fine tuning. I think the answer is somewhere in the human technology interaction that allows a human to bring creativity and imagination in places where perhaps the algorithm isn't focused and we'll see how that plays out. I don't think we have a perfect answer to that question yet. Natasha, I don't know if you have anything to add that there.

No, no, I think that's, I think that kind of covers it. We have another question, Joe, that I'd love for you to kind of take.

Okay.

Which is what makes you optimistic that we as leaders are up to the task of creating Ethical business once we are deep into the agentic world.

Wow, that is a really interesting statement. And I, I would say I probably vacillate between being really optimistic and being really pessimistic about it. And this is why I think my, the conversations I've been trying to have are with the most seasoned people in every, every part of the work environment. Because I, I do believe those, those people that have experience need to really bring it to bear right now, particularly given this question. So what makes me optimistic is I, I will say I haven't had a conversation where it hasn't come up. You know, there hasn't been a meeting with a client or a category or a business and we've been working with a lot of very large Fortune 500 companies, medium sized companies, you name it. There hasn't been a conversation where it hasn't come up, where you haven't gone somewhere where the conversation says, well, wow, that's a really interesting future, but I'm not sure we should go there. So I'm hoping it's on the forefront. But the big piece that makes me, I guess, the most optimistic is I feel like in a way this agentic future is going to give a lot of power to consumers in many ways. And they're going to create more and more demands for companies and their actions and they are going to punish bad behavior more and more. And it's not that I think consumers have to be the watchdogs. I hope companies really behave in the most positive way moving forward. But I do think in the culture and times that we're living in, we share that responsibility. Responsibility. And the more and more transparency that exists in the market, the more everyone's going to be under pressure to really, really perform. And I would imagine companies that do don't do a good job here will ultimately be punished by the people who buy their products.

I think the one thing to add to that, and I guess it's more of a call to action, and it's something that we've been speaking a lot about and have recently done a bit of research on, is the topic of KPIs and our incentives. I think maybe this is more of a pessimistic view, but we really need to make sure that we think of new KPIs or different KPIs in order to help us make the future more ethical and to a future that we want to actually live in. We are seeing in some instances that the same KPIs are being used. But actually as we're reimagining what the future looks like. We should also be reimagining what our metrics are as well that obviously incorporate AI. So it's more of a call to action to rethink what are the kind of KPIs that drive our business and what service are those KPIs and incentives doing that will help us build a future that we want to live in. Really great.

I think we have time for one more question. I see there's a question in here around is the mega efficiency alone motivating clients? I think I would say the reason anyone has gotten their feet wet in generative AI in their existing business for the most part has been about doing things faster, cheaper, with fewer resources. So I think efficiency has been the initial motivation.

But.

But we don't feel like that should be the motivation moving forward. It should really be around again doing things that were previously impossible because you now have the ability to marry human intuition, creativity, innovative thinking with a lot of really good skill sets that could come from an agentic system. So I think we're getting past efficiency as the reason to be thinking about this for your business. Any Natasha, there's one quick question. Can you give an example of a new KPI?

I was typing it in. It's okay, I won't, I won't send it. Basically there's a lot of kind of research that we've done into this. I can give you more kind of category level insights, Kathy. So you can really think of the relationship between your AI models or your innovation engines or whatever it is, or your agents, for example, and the impacts that it has on consumers. There's like a wealth of KPIs within that. So you can think of new categories that AI kind of touches. For example. Another example would be AI and sustainability. Obviously when we are expanding and incorporating generative AI, it obviously has a negative impact on the environment. So a potential way to think about how you offset that is making your back end a lot more efficient. So are you more efficient in the back end and does that offset the cost of actually having AI? So kind of things like that. Super high level thoughts Kathy, but I think that might kind of direct you in specific ways.

Great, excellent. Well, I think we are out of time everyone. You know, thanks so much for being such an active audience. We hope we've sparked your imagination a little bit. Feel free to reach out to either of us. Our email addresses are right here if you'd like to learn more or have additional questions. And again, for everyone who asked a technology specific question. Please join the next webinar given by Amir. I think it will answer a lot of those questions and be another facet of this conversation that's so important. Thanks again. Have a wonderful day, everybody.

Thank you. Bye.