Webinar: AI-powered consumer insights and opportunity prioritization

Good morning. Good afternoon. Depending where you are.

Welcome, everyone. How are you doing? It's a becoming evening here in Antwerp, Belgium. Would love to hear where you're calling in from. I'm seeing Tampa. Good afternoon, Katie. Hi, Carlos.

Look, we have an Amsterdam that's close to me.

Brussels, yes. Do we have an Antwerp somewhere?

Actually? Marek.

Nice to meet you. We found someone else from Antwerp.

Nice. All right, people slowly jumping in. So in a minute we'll get started as well. All right, let's get going. Yeah. So we're happy to welcome you today to talk about this days of subjective opportunity, prioritization that are actually over. You'll probably expect us to tell you how we gather insights AI powered today as well. And yes, we will as well. But first we'll actually focus on how we are disrupting our own process, coming a long way from using AI in Google sheets before and to where we are today. So happy to jump in on that. Yeah. So we're Board of Innovation, we're the AI Innovation consultancy and we help companies reinvent how they grow. And I say companies. So this could be companies across industries. Here's a couple of logos from companies we work with actually across consumer goods, industrials, more healthcare and everything in between. B2B and B2C are part of what we work with. There's roughly three buckets of work we do. So there's strategy on one hand, where we touch a lot upon today on both the AI strategy and innovation strategy on product, so how we actually built that and transformation, which is more around the operating model, capacity or capability building around AI as well. So yeah, in each of them we'll. We'll touch upon them a bit around insights, prioritization, empathy, creating opportunities, validation, et cetera. But we'll walk you through that. I'm Robin, I'm calling in from Lelystad close to Amsterdam. So happy to see quite some people in the chat from Amsterdam as well. And Francis is joining me today.

Yeah, hi. Just like Robin, lead strategist, pinheadboy for many years and calling in from Antwerp, Belgium. Maybe before we get started, there's the autonomous event that we're preparing for for the third time, actually, that we would like to invite all of you to. It's the world's largest summit for AI innovators. And given that you're here in this call with us, we assume you're going to be very interested in the two days, the 50 speakers. And we actually just crossed the bar of 10,000 registered attendees for the event already. So, yeah, check that out. We'll have some more information on that a little bit later on. So why are we actually reinventing this innovation process? And it's a bit of an open door with just having the slide for the autonomous event. Right, but the reason we're rethinking a lot of this is because of the belief we hold that our world will become more autonomous. So actually after the digital transformation, we see autonomous as being the next big wave of transformation. And we'll see that all around us. Right? So that doesn't just mean us typing things into ChatGPT, but actually we will start to interact with many different interfaces. Some of them will be digital, some of them will be products as well. Think about self driving cars for example, but also different apps and interfaces written just for you, like personalized applications, even businesses being run by AI at some point. It's a lofty goal and a lovely or a far out way of seeing things with AI. But it's enabled or unlocked by this realization that the cost of anything digital will actually near zero. So following this exponential view from Azim Azhar, you already see that storage and bandwidth and CPU power dynamic, they've all decreased in cost exponentially. The one thing that still was a quite large expense was the engineers to write actually all that software. And if there's one thing we're noticing AI or LLMs rather are really good at is it's coding. And Currently, even though 50% of code is already being written by AI, currently a lot of that code is still boilerplate. But over time this will mean that together with these other decreased costs that everything. There we go. Everything digital will be generated in no time at no cost. And we need to be ready to innovate in such a world. Right, because innovation will look very different in such a world. Not to completely redo everything we talked about in our previous webinar, but for completeness sake, we do want to highlighted. So we see AI adoption follow three waves where wave one focuses on time and efficiency, wave two on quality and better output. And in the previous webinar we talked a lot about transformation, right? So enabling new business models with AI. It's not necessarily true that wave one has to happen first to enable wave two and three, you can see them as separate. And in this webinar we're mainly going to talk about wave one. This is going to be our focus today with some practical examples. And that's because most organizations will actually focus here to make their current processes more efficient, to save time, to save cost, which makes Sense. Right. Because this is closest to home. So just using AI in our innovation approach, that's not something we want to do quite that ad hoc. We do want to reason a little bit and think, okay, what are the pain points we've noticed over the years? Right. Where can we improve? And I'm sure that we all have war stories about challenges we've encountered in innovation. So I'm sure a couple of these will ring true for all of us. For example, entire projects being based on anecdotal quotes that someone heard someone say and that stuck with them debating for months, what's the perfect concept? Instead of actually going out the door and validating a few of them with consumers prioritizing based on subjective hunches or own opinions, more time being spent prioritizing rather than actually chasing opportunities. And it makes sense, right? Because we only have limited resources, so it makes sense that we want to be sure about the opportunities we chase. But after doing this for quite a few years, we do notice that a lot of time is being spent on prioritizing rather than actually going out and chasing opportunities. We've actually tend to, we've actually lost a little bit our bias towards action in hesitating toward, in building and showing progress. So in chat, feel free to leave a couple of the other things you notice. Let us know if this is resonating with you or not, while I move on to the methodologies that we're using. And we're not married to any one of them, but over the years we have seen that using a combination of design thinking on one hand and lean on the other gives us good results. Lean without design thinking lacks the empathy. Design thinking without lean likes a bias towards action. But in using both approaches and both methodologies, we do get a nice mix of things. Now what we're asking ourselves in using AI is how does this balance change? How can we still be very lean and maybe even more lean and be faster and do more without losing that empathy? That's a very crucial question we've been asking ourselves. So let's dissect our own innovation approach and look at a couple things there before we actually go to some of the use cases we found for AI.

Yes. So I will actually dive into what we call strategic opportunity spaces to narrate this a bit further. So you might have seen this double diamond before. For I think the last two decades, this has actually been the main model being used across the design thinking practices where we always diverge and converge. So first we focus on the problem. We get our insights and signals and go broad there and then cluster and dissect and prioritize opportunities to focus on. And once the idea is clear, we'll actually start diverging on the solution. So we generate ideas for the selected opportunities and iteratively validate them, narrowing down to the final solution that we look for. But we actually used to think that this was exploring lots of ideas. So this is an example that we're going to use throughout this webinar. But we'll refer to a case we worked on with Nespresso. There we identified 12 key opportunity spaces for them and we focused on two of them to explore in depth, actually. So that's where we narrowed down our problem. And then we created just over 100 value propositions for those two opportunity spaces. We narrowed down again through iterative validation to 50 ideas selected and five final validated concepts, plus four that we're ready to pivot to put into the innovation pipeline there that year. And the idea was to reuse the other opportunity spaces over the years to come. But yeah, what you see in there is that we actually are very heavily limited in the amount of opportunities we can explore because we identified these 12 opportunity spaces that they could play in and only one out of six was actually focused on. So at the gate, or even in front of the gate, we already had to prioritize what to focus on. Now imagine if we could actually bet on all strategic opportunities. It could mean that we can spread our bets because all opportunities are there for a reason. It's not that one is good, the other is not. Of course there's a prioritization happening to make choices, but opportunities are a strategic opportunity for a company for a reason. But yeah, today with the way we've worked in the recent years, we aren't able to design multiple value propositions, 100 plus propositions per opportunity space in that case. So that doesn't work the way we want. So what have we asked ourselves? Do we actually still need opportunity prioritization if we work AI powered? We don't think we need to, and we're going to show you in a minute. And do we actually still need to define opportunities instead of moving straight into value proposition design? Yes, definitely, because the strategic opportunity is more than just a customer problem to solve, and I'll show you why. So strategic opportunities only exist when there's a big unfulfilled customer need or wish. This is often what we focus on. What makes it attractive is this popular, frequent, and is there an intensity to the pain or gain that the consumer or end user has that we could fulfill, but it's also relevant to your company. So is this consumption occasion, the target audience, the brand fit, are they all relevant to your company? And do we believe that we can beat what's out there? So do we have a right to win? Can we actually produce this? Do we have a competitive advantage? Or could we have. And only these three things combined make a strategic opportunity? So it's definitely more than just a problem to solve that makes it actionable. So how do we actually identify strategic opportunities? We can't predict the future, but we can define what plausible features are based on patterns in trends and signals. And when we map those plausible futures to hedge our bets, we make sure that we don't create tunnel vision. And we look broadly in the future to identify opportunities, always specifically for us as a company. Now, if you would make a cross section of that funnel you just saw at the moment that your innovation horizon is on, you would see something like this. So Here we took 2030 as our target horizon, and then all those plausible futures have a scenario in play. And yeah, these plausible futures describe the impact of the signals of change on the world around us and on us as a company, but only when we identify that future to be relevant to us. We dissect opportunity spaces that are actionable for us to develop solutions. Opportunity spaces help us build a bridge between the fuzzy market signals and concrete concepts. So if you look at it on a scale of concreteness, where we want to get is how to win. So what are the pains we need to resolve? What are the value propositions we could build? But where we get our information from is where, yeah, where to look. So what are trends or signals we see, what are future scenarios we map, and in between is where we want to play. So which jobs have we done do we resolve? And that's the strategic opportunities we look for. So that middle ground now to make choices between those sizing is actually crucial. Francis already touched upon it a bit. Yeah, we often see very subjective discussions about choices at this stage in the process. And we try our best to actually be objectively sizing opportunities as much as we can. And these are just a couple of examples of the way we do that. So we prioritize jobs to be done where we look for those most important underserved spaces based on consumer data, some consumer feedback. In experiments, there's also ways where we size the opportunity by identifying share of throat or stomach that we might be able to steal. So then we look beyond just our space and there's also possibilities to size for use case. So in this case with loop making earplugs, we actually sized the potential new use cases for them and clustered those into opportunities to focus on. So we were able to size those industries that we could play in to see where the most potential is to then focus on. So on one hand, an opportunity space is that bridge between the fuzzy market signals and concrete concepts. On the other hand, it's a prioritization tool to help us focus our innovation efforts. But what we see now in an AI powered world is that we don't need that second version, second hand anymore. If you go back to the questions we asked ourselves, do we still need opportunity prioritization if we work AI powered? No. But do we still need to define opportunities to concretely build value propositions upon? Yes. So that's what we still focus on there. But the unlock is the ability to explore all identified opportunities. We don't need to prioritize them anymore before even exploring them. So there's no need to prioritize in front of the gate yet. But first get into Wonderland before we actually figure out what's promising there. Yes or no?

Yeah, it's a huge unlock. So what clicks for me is that the focus and clarity that opportunity spaces bring is still crucial. Do still need to size and select to get to an opportunity. But once it's an opportunity, why the hell would you not research it? It's so obvious now that you can finally research all opportunities that you actually would do it. And to do so you need some exponential thinking. That AI allows us to do so. For those who've already joined the webinars on the Stingray model, this will look very familiar. But also for everyone familiar with design thinking, you'll see that this is a slightly changed version of that, where there's much more emphasis on the development, iteration and testing phases. Which doesn't mean that empathy isn't important anymore. It still is, very much so. But the balance has shifted from the beginning towards more the iteration and test phases, because insights have become very cheap and easy to get. Like, we can pick up all of those signals quite easily. More on that later. So that enables us to quite quickly develop a whole range of exponential concepts. And yeah, we talked so much about opportunity spaces because that is actually the starting point for, for this whole exponential process. So when we transform our opportunities into a lot of sacrificial concepts, we can actually get to 80% of validated ideas in weeks. So let's collectively stop wasting money on subjective debates. About what are the best concepts or what are the opportunities to go for and rather start using sacrificial concepts instead. And this is actually not a new way of thinking. We've always fueled our way of working with new technologies and I'm sure we're not the only ones. Actually, a couple of years ago we had this vision of becoming half human, half machine. And that was a very far away vision as well. But it became reality much sooner than we expected. So maybe to dive into the things we did in the past. For example, as you see in Empathy, we ran and recorded interviews digitally and put them in Dovetail, a research platform I'm sure lots of you are using as well to tag our interviews, to cluster the insights and to get to good Personas. Partly due to Covid, we started using Miro and we started collaborating with many more stakeholders together to get more complete ideation sessions. And in validation, we're still running ad campaigns as AB tests, we're running digital storefronts, we're doing value prop surveys. So we did do a lot of things already. But what you see is across all these three things, gathering feedback, creating concepts, gathering insights. This is all being disrupted with AI insights. They're becoming cheap, or at least the first insights to start from. So how can we now leverage this? Let's start comparing the old way of doing things with the new way. We used to think that this was actually rigorously iterating and validating at scale.

Yeah, so you do actually recognize the numbers probably from before. So this is the way we used to do stuff that we thought was at scale. So we identified opportunity spaces based on social listening, historical data and search terms, search listening, etc. And then actually dove deeper into those specific opportunities page. We selected and ran Empathy. And here we already thought we were smart and ran things like video asks. So we actually interviewed people at distance, we asked them to answer specific questions with video showing us how they consumed coffee in this case for instance. And then we always got together with a team of experts to do needs based ideation. Like two full days of workshopping in this case. You can imagine the opportunity cost of bringing such a team together and then actually prioritizing through consumer connections. So running idea prioritization experiments to narrow down and then actually going through multiple validation rounds. So we ran more focus groups online where concepts were being we being assessed by consumers in store testing, actually you see me on my back right there in the picture. And other similar conjoined experiments, you name it, were part of this iterative validation process in that case, to get to your pipeline of ideas in the end.

Yeah. And if you imagine this blue line in the middle as a scale, you, you could imagine that the weight is dipped pretty heavily towards the left side of things. Right. We spend a lot of our resources and time on the beginning part of this process. While actually the value, most of the value is in actually going to consumers to their top right image, where Robert and a colleague went in store and actually put prototypes in the hands of people and got real feedback and were able to iterate from there on out. So the more we can do to spend more time with consumers to actually prototype and iterate with them, the more we should, and we should take that chance. So let's have a bit of a look at how we disrupt our own way of working nowadays on the empathy side of things. We were already using social listening, but when we were doing so, we were doing it in a very manual way so we couldn't look at a whole lot. Which I saw some things pop up regarding bias and the fear of being biased with AI. Well, actually when we were doing social listening in the past, we were also quite biased because we were so constrained on which things we could look at. It very quickly becomes cherry picking. Now when we're doing agentic social listening, we're actually going through thousands of social media posts. We're getting agents to analyze the pains, the actions that people are taking in each one of them, clustering that and summarizing all of that. And you could argue that's a more objective way of at least doing social listening and getting those initial empathy insights. AI moderated interviews, I'm sure is something lots of you are familiar with. We've been doing that for over a year now. But also on ideation, maybe to jump.

In with one of the questions on the AI moderated interviews as well, because Alexander is mentioning the losing touch on ethnography, for instance, we also use this in two ways, so also in a hybrid approach. So you could also see it as a way to do interviews at scale and get quick qualitative feedback very quickly without having to run interviews or yourself and then actually diving deeper with the most interesting consumers or respondents you found. So there's also. It doesn't mean that we replace everything human with AI. There's also a hybrid approach in between. But especially finding that optimum of where the hybrid optimum lies is going to help a lot to do ethnography at bigger scale as well, while not losing touch.

Yeah, thanks for jumping in there. And maybe to add even one more point there for Alexander, see this as another approach, another tool in your tool belt. Like we were saying in the beginning, we are methodology agnostic. This is another way of running things and running projects that we're seeing is a very good fit in some industries for some projects. Obviously it's not a fit for everything. But more on that later. I was talking about concept generation and synthetic Personas. We also have a bit of a demo there at the end, where we already have built a module that can generate concepts in a matter of minutes and actually validate and iterate on those with synthetic Personas. And then finally, also on validation, we're making strides where on desirability, we're already seeing some good initial results, but we want to move forward, obviously, and we also want to get some initial validation on the viability and feasibility side of things. So let's talk through these. Right? And while we're talking through these, we'd love to see in chat how you've been using AI and if these use cases resonate with you or if there's any concerns or doubts that you're feeling while we're talking through this. Right, because we certainly had those when we were first using them and also still have them for some of these things. So this is about social listening, which I was talking about earlier, where in the past we were manually sifting through on the order of hundreds of mentions per week, where right now, in a matter of days, we're analyzing thousands of posts and clustering those, not only actually with AI, but once we have a whole table of pain points and actions that people are taking. We're also using ML models that have existed for a longer time to cluster them into dendrograms to make sure that the pain points we're clustering do fit together.

And taking the example from the empathy interviews, again, you see me here running a project where we were developing a new user experience for a potential new coffee app, and it actually took a couple of weeks to get qualitative feedback on 10 concepts in this case. And what we can do now, think about that hybrid approach again. We can get qualitative feedback on 30 or so concepts overnight, even during a workshop where we would develop new concepts overnight, get the feedback in, and then actually iterate on it again the next day. So this would never be the final step of validation, we believe, but you could easily run a couple of these overnight feedback moments in a week of workshopping, for instance, and be at that 80% level that we mentioned, instead of needing to be first time. Right. Once you get into consumer feedback mode.

Yeah. And finally on ideation, in the past, we spent days or sometimes weeks with quite a few people to get to a whole bunch of ideas. What you typically see there is one. It takes a lot of time and a lot of cost or opportunity cost to get to that point. But also sometimes you get to some sort of consensus where you get to a lot of ideas. But they're all rather similar where right now we're detaching from that way of thinking of getting to the perfect concept. And it's more about how can we get to different concepts that are each sacrificial and each, in their way, test a certain assumption. So it's more about how do we test our assumptions rather than creating the perfect concepts. So that's some of the use cases that we are using. But like we were saying, we're definitely not 100% clear about this whole process. We're not 100% sure about everything here. So maybe let's talk through a couple questions that we have here. And if you want to add to them, feel free to maybe first. Robin, will this replace my job?

No, it won't. But as innovators, we move further down the process. So now we are actually, the creators were building concepts before we start validating them, and we can actually become that curator. So we can have these systems help us generate concepts. And yeah, basically we have a gift to be able to work on concepts with a higher chance of success rate. In that case, we let the first hard, dirty work be done before our brain power comes in again. But they won't replace us.

Yeah, I'm actually quite optimistic about this. The more time I can actually put cardboard prototypes into people's hands again, or pictures in front of people, that would make me really excited, actually. Can I trust synthetic validation? We've done actually quite some testing here ourselves and with partners. And there's a yes and a no to this. So before we go into both of them, you could ask yourself, do you need to fully trust synthetic validation? Because we're setting quite with this question, we're immediately setting quite a high bar for synthetic validation. Basically, you're asking, is AI at the point where it can completely mimic humans? Which it's not. But what we're seeing is it is good enough to give initial feedback on ideas, to iterate on them, to get to that point where a sacrificial concept is good enough to start putting it in front of people, where we're not seeing it as Being good enough yet is where it's replacing one to one survey testing or survey validation with actual people yet. But there's another side of the coin there as well. Maybe. Robin, you want to touch on some of the research here?

Yeah, yeah. So on the, on the left you actually see a recently done example where ChatGPT is not able to be as random as humans yet. We're not at that point yet. And those graphs actually show ChatGPT giving a random number and between 0 and 100 for a thousand times. And on the right side is the company checkbox that ran this with actual human. And then you see that, yeah, we're, we're actually more random than ChatGPT is. But at the same time, Ethan Mollick, probably known to many of you as one of the masters in AI, he actually is already predicting that at some point online questionnaires are becoming useless or obsolete because we simply can't identify whether it's AI filling it in or humans anymore. So, yeah, there's also an optimum balance to find between the two. And we're sort of forced to actually look at both sides of the coins as well. And this is actually a question we asked ourselves a couple of times, or actually one of our colleagues, eto, challenged us with why does innovation have to be faster? What do you think, Francis?

It's a good question. It doesn't have to be faster or as fast as we can do with AI necessarily. It's just one of the unlocks. Right. And I think it helps to think in terms of the three waves here. So one of the waves that we can use AI for and that it unlocks is to do things faster and more cost effective. So there, of course, we do need to take those gains. We have touched also on the argument that it allows us to spend more time later in the process where our time has more value. But I think it's also crucial to think about other things that AI unlocks, and that's about entire new business models. So this is just a starting point for us to experiment with AI and innovation. But for example, one of the, the great things we've seen AI help us with is to enable business models around personalization. So when AI is super fast, when it allows you to in seconds come up with who's writing something in chat? Victor, the milder. If it allows us to create like a piece of clothing or some food for Victor and for him alone, then we enable business models that focus only around personalization and hyper personalization, which is tremendously exciting. So it doesn't have to be, but it does save us on cost and time. It does allow us to focus on parts of the process where our time has more value. And finally, it allows us to enable new business models.

Yeah. So why are we actually changing our process? There's a couple of things. A higher success rate. So with early stage opportunity prioritization, we're forced to opt for the, at that stage, most promising opportunities, which are often the most safe bets. And if we put that prioritization at a later step in the process, we can simply go broader, look at more fringe ideas as well, and get a higher success rate on those. More personalization. So high speed iteration also enables us to ideate on those fringes and get to ideas. Like oxagrip for instance, which was a concept or product that was designed for our tritis patients, but has a broad customer base today. And without the ability to focus on that fringe, it would have never been marketed. And yeah, we're able to focus on impact rather than coming up with concepts in this case. So that's more related to ourselves in the process.

Yeah. And then a quick question from Christian sun in the meantime, what does higher success rate mean in this case? Well, it's, I'm sure you're familiar with it. Lots of the concepts we test initially, they fail. They don't meet market demand. It's not really solving the consumer need in a way we thought and a way we're trying to tackle. This is, for example, in a concept generation module, we're not only coming up with concepts and we're not only generating ideas, but we're also having all of those ideas tested. We're having different sets of AI agents give feedback on them, and then we iterate each of those concepts based on the feedback that we're getting from them. So that iteration loop that already happens, that enables us to have a higher chance of success that we're solving a consumer need when we actually put those in front of consumers. Does this work in my industry, which is a great question. I believe we're seeing it works in more and more industries as we go along. Where we've proven it works is in cpg, in fashion and in furniture. And actually at the moment, we're also doing multiple projects in pharma. So in more and more industries, it's starting to work. One thing to keep in mind is that we are mainly working with language models and they're very good at writing out text. Right. So they're good at coming up with ideas, at analyzing things. At summarizing things. So it's great to come up with different concepts aided by visual models. But for engineering, for example, we've seen that it's not so good yet in developing industrial concepts. Maybe skipping through a couple of these questions for the sake of time, but you'll definitely get these slides.

Yeah, we definitely want to tell you a bit about our vision, so how we actually help the most ambitious businesses to implement this approach as we speak. And for that we have to reframe how we currently think about AI, about innovators and our roles in that as well, a bit. So let's just compare today with what we see happening in the next years. First of all, AI is a tool that humans use today. We actually believe it becomes the connective tissue that allows humans, data and robots to collaborate over time. We touched upon the next point a bit already as well, that we are creators today, creating stuff from scratch. And especially I have a design background. I love creating designs from scratch. But we're becoming editors or curators slowly, where we have the initial product being created and critique it. We also become the manager of AI native workflows and products rather than people and processes. So the whole process is going to change and coming back to that success rate a bit that was already discussed and asked about. This is basically what you should envision compared to the initial approach. The example of Nespresso we've shown now, we could actually create hundreds of value propositions per opportunity space that we identify. We don't have to choose first. We create all these propositions and then iteratively start validating. And the darker the blue in this stingray gets, the more human involvement is there. So you just slowly go from synthetic validation to hybrid to human validation, undesirability, viability and feasibility. So we would always build agents that are built on either novel data or company data on production capabilities, for instance, or financial models that are already known. And a hybrid version of the two.

Maybe to add to this. Again, Christian, I seem to be answering your questions, mainly asking, is it adaptable for B2B? Yes. But you have to ask, does it make sense for your team? So for some B2B teams, it does make sense to go for 500 value props and indeed look at 20 different opportunity spaces. But also for a lot of other B2B teams, they don't want or need to go this broad. And then it doesn't make sense to set up and fine tune a whole AI process, because this is not just us typing things into ChatGPT, right? It takes time to set up something like this so that it actually produces something of value. So it doesn't make sense for every B2B team to work like this. Great question.

Yeah. So maybe to summarize a little bit, what does that AI forward process actually enable us to do? Well, to work harder. We increase our robustness by running more iterative validation. It's becoming better through preventing prioritization at the front of the gate and actually developing more concepts simultaneously. Faster through shorter iteration loops and stronger by being more cost efficient in validation. Starting synthetic first before going into expensive experimentation and the be always on approach. And we'll actually show that in a bit soon. But do you actually remember this one?

So these are the use cases we're talking about. Right? These are the things we're actively doing now with AI in our process. To show you a little bit of the vision of turning this into an autonomous process, we see a future where we're combining these use cases into different modules. We could, for example, have an insights engine that that makes use of not only social listening, but also proprietary data pools and scientific breakthroughs by analyzing patents. For example, we could have a concept engine that's being supported by a validation or simulation engine, something that already builds the first designs, and finally something that scales. So this is all a little bit abstract, but let's look at maybe one piece of the puzzle where we've built our own QWIK concept generator. And I can actually show you a really quick demo here. So let me move here. Hope you're seeing my screen. So this is our own proprietary engine that we've built. This doesn't go or take in more data than what I've written here. So, for example, this is for a project called Pasta Opportunities for Gen Z. This is for Robin's new pasta brand that he doesn't know he's going to launch yet, but he will after seeing the opportunities here. We filled in a couple things on project description and once we've saved this project, this tool is going out online. It's going to find relevant trends for this project. It's also going to look in a couple trend documents to find initial themes. And once this runs, it takes a couple minutes. So I already ran this. It's finding themes online and turning those into opportunities. For example, opportunity around customization. This is around using digital platforms. And in each of these opportunities you can explore the ideas. What I did is I actually know Robin really likes music, so I asked for a focus on combining pasta and music. And we're already getting more relevant themes. Right. So the Groove and cook pasta kits, for example, are really nice and then within it you can look for ideas and even turn them into concepts. So maybe to end with the concepts that you can then see, they get turned into full concepts. For example, the 3D pasta creations could really see something like this work. So this enables us to get to level of sacrificial concepts really quickly and then put it in the hands of people.

Yes. So it was just to show you a bit what we actually. Yeah. What we've been up to. We tried to answer a couple of questions in the chat for you, but we're close to the end of the webinar. If you have more in depth questions, feel free to actually send us directly as well. Also, don't forget, join us on the autonomous summit early next year. Actually in just over a month's time, Nico has shared the link to register in the chat. It's pinned. But beyond that, feel free to reach out to us. We're happy to share more in depth with you, exchange thoughts as well. This was just the tip of the iceberg of how we've been changing our process. Both the AI powered way we work and the more autonomous way we work are part of today's project as well. So happy to share more there for those who asked. The recording will be available if you've registered and we'll share the slides as well for that. Happy to have had you here and wish you a great evening or a great rest of your day if you're on the the other side of the pond for us.

Thank you for joining us. Today was great having you speak.