Webinar: the Executive Blueprint for Autonomous Innovation with Philippe De Ridder

Hello everyone, and welcome to today's webinar on the Executive Blueprint for Autonomous Innovation. I'm Philip derider, Founder and CEO at Board of Innovation, and super nice to see you all here today, gathered on our next piece of thinking around autonomous innovation and especially geared towards more executive thinking around it. And more specifically, I want to take you on the journey of on how to become a leader in an increasingly autonomous world. If you have any questions in the meantime, feel free to drop them in the chat. My colleague Vincent will be monitoring the chat, answering any questions on the go, and we will also be taking questions at the end of the webinar, so stay tuned for that. Let me start with introducing you to Devin. Devin is the world's first AI software engineer that Cognition Labs introduced last week, and you might have seen a couple of things going around on that. Ethan Molnik asked the Devin AI agent to go on Reddit and take website building requests. So he just asked that in the chat to Devin and that's exactly what it did. So it served to Reddit, it filled in the captcha form, it created an account, it looked for the right Reddit form thread, it started posting, it started monitoring answers, it actually even asked for a price. So it gave a budget rate of 50 to $100 an hour for its building ask. So just to point out, Devin is just an AI agent, but slightly more advanced. In addition, Devin can also autonomously build and deploy apps end to end. It can learn how to build something just from a blog post. It can start doing real jobs on upwork for the people that have tested it 100% autonomously. Here on the right side you see Devin's workspace, and this was based on Ethan's request. So it created its own plan for that request. So it said, well, I first need to log on to Reddit, create a post in a subreddit, I need to edit the post to comply with subreddit rules, I need to attempt to resolve initial posting issues, etc. I need to monitor the answers to that, etc. Etc. And it adapts its plan as and when you collaborate with it. A quote that really summarizes the experience of working with Devin is that it's less like an assistant helping with code, like the copilot, but more like a real coworker that's doing its own thing. And it feels very different because it's an autonomous system. At around the same time over the past week, there was obviously Nvidia's big conference, which I'm sure many of us have watched. And there was their CEO on stage making the claim that everything that moves in the future will be robotic. I'm sure that he didn't intend to include humans or animals in that statement, or at least I would hope so, but you never know. He's obviously been quite right about quite a couple of things with Nvidia over the last couple of months. And they presented their new model called General Robotics 003, or just in brief Groot, for the fans of Guardians of the of the Galaxy. And it's an entirely new model that will help robots to learn based on how humans behave and basically learn both in simulated environments much faster than they used to, as well as in real environments. So just two examples of what just happened in the last week. So, on the one hand, in terms of software encoding, a new autonomous agents, as well as a major leap forward around autonomous robots going forward now, zooming out, where should we place this all in a historical context? And what does that mean for all of us going forward? I want to take you just a little bit on the journey of an evolution in information technologies throughout history. So the first information technology was writing, which we invented around 3,000, before Christ. And basically the primary function of writing was that we could record or document information and hence preserve that for next generations, etc. About 3,000 years later, we invented the printing press, which allowed us not just to record information, but also distribute it. It was our first more mass medium to actually distribute information across the world. And whereas writing had unlocked more complex forms of society and also the first forms of more complex economies, because we could actually keep books and transactions, and first accounting systems arose from that, printing allowed us to get into or was a key enabler of first the scientific revolution and then the industrial evolution, which obviously changed the world as we know it today. Then in the 1990s, coming closer to where we are today, the Internet came to exist or became more mainstream and search as a component of that, which basically maximized the capability of distribution, distributing knowledge and information to the whole world. And if you just figure on how that has changed our society and our industry, the fact that we're having this very webinar right now is obviously because of the Internet. Just remember where you've potentially last purchased your last pieces of clothing or food items, etc, where you read your news, a lot has changed. And today's most valuable companies around the world came to exist because of the Internet with the likes of Google, Meta, Amazon, etc. So just to recap of where we are Right now, basically, information technologies that have offered us the ability to record information and to distribute it around the globe. In the last year, we've started to see the next capability, not just recording and distributing information, but information technology that actually generates new information with the likes of ChatGPT, Gemini, Sora, et cetera, generating new information in the form of text, in the form of audio, in the form of videos, with Sora, in the form of images, with the likes of Dall? E, Adobe Firefly and others. So that was 2023 and we've only scratched the surface of how that is actually going to change the way we live, the way we work, the way our economies work. And we're already at the eve of the next step in the evolution of information technologies, which is towards autonomous AI agents that not just generate new information like text, videos and images, but that are able to take action action in the digital world with the likes of Devin the Firl Dub, the world's first AI engineer, as we've shown you the example, as well as the model developed by Nvidia, which will get us to more advanced robots, to actually taking action in the real world, in the physical world as well as in the digital world. So while we're still trying to figure out how generative AI is will be changing about everything we do, we're already at the verge of the next wave in this and we're basically adding two new capabilities in just one year. Technologies that generate new information as well as that will be able to autonomously take action based on any assignment that we've given them. So this will likely be the biggest transformation ever of our world as we know it. And also just acknowledging that this will start if, if you are starting to have an existential crisis about this at some point in time, that probably means that it's a good indicator that you've gotten what AI is doing, because it will be quite transformational and in some ways a bit scary as well, if you think about the likes of Devin and also what they could do in terms of misinformation and other elements. So what we're thinking about this evolution in information technology and how that's going to change about everything that we do in order to seize that opportunity. As an executive, there's three things that we need to look at. First, we need to envision that future, envisioning an increasingly autonomous world and how we will thrive within that, rather than starting from the as is from the current state. Secondly, we need to be building new autonomous engines for always innovation and growth. And a bit more on that in a minute. And thirdly, we will need to look at transforming our operating model rather than looking at use cases, function by function, and basically only looking at things through an efficiency lens. So we need to envision, build and transform. And I want to take you on a little bit on that journey throughout today's webinar. Starting with envisioning. For us, we're envisioning a world where everything is becoming increasingly autonomous. What are we talking about? So envision autonomous assistance. In the not so near, in the, in the near future story, we will likely all have personal AI assistants that know our personal preferences, that know our preferences in terms of the clothing we would like to wear, what fits the color of our hair, what we would like to eat, what, what our goals are in terms of being healthy, of doing sports, etc. And they will start curating options in terms of products that we might want to buy, not want to buy, in terms of daily routines that we might want to adopt, et cetera. And increasingly as companies making services and products that are geared towards users and consumers, we will need to take into account to actually design them as well. So they work for the AI assistants that that we will all be having. And at first they will just curate options, but as just shown with the likes of Devin, if and when we allow them, they will start to take action on our behalf. And not just suggesting three good restaurants to look at, but actually booking a restaurant, buying clothes or whatever it is. We're also talking about autonomous products. Think about self generating user interfaces rather than fixed websites or apps. We're currently looking at a technology that basically self evolves a website and that alternates where buttons are call to actions, etc, learns from that rather than going through traditional user research, prototyping, etc. A self evolving product. But also think about autonomous products like in the space of robotics or autonomous cars, autonomous agricultural vehicles, autonomous delivery, etc. Autonomous interfaces, as already just said, self evolving website, even self generating interfaces. We've seen a couple of interesting demos from Gemini, for example. We're talking about autonomous agents, where Devin I think is the latest state of the art example of. But we're also talking about autonomous processes to get to, for example, new products or new services that we will be launching to market. So this is by no means an exhaustive list, but just to give you a flavor of what we mean by to live in a world where increasingly things will become more autonomous. Having said that, within that future vision is that the current focus on leveraging AI for efficiency is by no means capturing the full potential of AI. AI is not about writing our emails a bit faster or a bit better. It's going to be about completely reimagining the world that we live in, as well as how we operate within that as a business, as well as each of ourselves as individuals within that. So within that strategic framing, I want to take you to what we see as the strategic options to capture the opportunity of autonomous innovation. Beyond looking at efficiency gains and productivity gains, I want to take you to this overview which will list on the vertical axis, what we deliver to the market, so the products, services, experiences that we deliver to the market, and on the horizontal axis, how we get to make them, and basically the operating model that underpins them. Splitting that out in a bit more detail, starting on the vertical axis with leveraging AI for the existing types of products and services, and operating that within the existing operating model. I would say that's generally where most of the current focus and investment is going into, which is productivity upgrades to either existing products or the existing operating model, with the key targeted benefits being cost efficiency. While there is obviously value in doing that, we want to expand the horizon and we think that's part of by no means unlocking all of the value that AI and autonomous offers. So on the side of the how we are organized, we think the discussion should be much more around envisioning what the AI native operating model for innovation is in the future. What we would call system innovation. How can we envision a new autonomous innovation engine that allows for market launches at exponential speed and success? What if, for example, in consumer goods we could go from identifying a new trend to having a new product launched in just five days rather than two years. We will never get to such a future operating model if that was something that we wanted to build by optimizing the existing parts in the current process, doing insights slightly faster, generating concepts with the help of GPT etc. You will need to rethink the entire operating model, including sourcing production, go to market, etc. The key benefits here being cost efficiency, driving revenue growth, but also especially market entry philosophy of being to market exponentially faster. Then if we extend the horizons on the what we deliver to market, I think the interesting opportunity is to envision what the AI native products and services could be if you were to think from a blank sheet of paper and what they would be in your industry. So we're talking about product innovation, launching AI powered products and services that would deliver a better customer experience, but that would still run within the current operating model, then if we were to combine both and be even bolder, we would be looking at breakthrough innovation which is launching entirely new AI native products and services operated by a new operating model that would disrupt existing market paradigms entirely. So the key benefits here being that they would either be industry creating or industry redefining and obviously being focused on revenue growth rather than cost efficiency. So within this overview and the four strategic options that we see to capture autonomous innovation, there is no right or wrong. I'm not advocating to just focus on one of the four. The advice is to build a deliberate portfolio approach across all four to have a clear vision on the future and then build that portfolio around that. And for that. Now, when we play out this matrix and when we look back at that last information technology breakthrough which was the Internet or digital, we see that most market shifts have been haven't been captured by incumbents, but rather by new entrants into each market space. So if we play out digitally native products and services and digitally native operating models, in this overview, we're thinking about the likes of Netflix in streaming, HelloFresh in meal delivery, Teladoc in telemedicine. And it's a pretty long list of new entrants and players entering new markets with a breakthrough innovation that redefined that industry or created entirely new verticals or industry. And now with AI as the next breakthrough information technology, there is a new opportunity to capture this time around rather than just only staying focused on the efficiency and productivity gains in that bottom left box. If we look at some early examples in terms of a more AI native operating model in food and beverage, we've seen Notco, who were also a speaker at our last conference back in December, and they've basically built up an AI native operating model from from scratch, looking at algorithms, matching ingredients, looking at consumer propositions and developing new food and beverage products from that. If we're thinking about product innovation, we've seen an incumbent like Adobe to respond relatively fast after the likes of Midjourney and Dall E developed text to image models and have been very successful with launching Adobe Firefly into their suite and with their enterprise clients. When we think about more breakthrough innovation that's truly changing or redefining an industry or creating a new one, we're thinking about the likes of the Humane AI Pin or the Rabbit device, both in consumer tech and making a play at what would be an AI native consumer tech device that could potentially replace our smartphones. Not to say that they will be successful and will be winning the market over from Apple and others. But they are making an AI native breakthrough play in those markets. And that's to me the interesting question as well for each of us to ask and have the internal discussions around with people within your companies. What would be the humane AI pin in our industry and is that an opportunity that we are interested in capturing? What would be those AI native products and services and what would be that AI native operating model for our company and our industry? And here I just want to pause for a second and based on that matrix, kind of get your input as well by pulling up a poll. There you go. You should see that pop up on your screens hopefully. Does that work? Votes are coming in. Sorry, that was a short delay. So the first question is where is your organization investing right now amongst those four options? And let me just maybe play them again on the screen. So is most of the investment going into productivity upgrades? Is it more going into system innovation, looking at AI native operating models, more into product innovation? Or is the investment going into breakthrough innovation? And the question is more oriented. If you were to look at it not for just your team or your division, but if you look enterprise wide from an executive lens, where is the investment going? So we have about 150, close to 200 votes coming in. I'll just give it one more second. There you go. And there are the votes coming in. So most of the investment is flowing into productivity upgrades. Then a very clear second product innovation. So AI native products and services and then system innovation and breakthrough innovation clearly take a backseat in terms of investments in companies. Let me open up a second one and this would be your personal perspective. Where do you think your organization should invest in the most if you had to prioritize only one of the four And I'm just going to open up that one as well so we can kind of compare them one to another. Sorry, just giving a second for the votes to come in here as well. I can already say they're different, but maybe it's also biased by potentially an audience of innovators in the virtual room. Right. It's quite an equal race. Let me just. Votes are still coming in, but just in light of time, let me pull it up here. So it's quite an equal race between system innovation. So looking at AI native operating models, breakthrough innovation, product innovation and much less the focus on productivity upgrades. So I think the, well, biggest difference or let's say the biggest similarity is probably in product innovation against what's happening right now and where you would see most of the investment going and then the biggest difference is across all the other three. All right, thanks a lot for joining in on that. I'm just going to pull up my slides again. There we are. All right, so I hope that's given you a bit of a strategic framework to think about the strategic options to capture autonomous innovation and think about a portfolio approach across those four. And hopefully as and when you have discussions internally around initiatives that are being launched in terms of AI transformation, product innovation, etc. It gives you a framework of where they fit within those and can help guide those discussions. Now, when we look at envisioning the future, and if you want to bring this upwards into your organizations, we need to be thinking about what the impact is of autonomous innovation in terms of what the business case is for the whole company or how it will impact market valuation. What you see on this slide is the breakdown of total shareholder returns, which basically breaks down between enterprise value and then a couple of financial influencers like buyback rates, dividend yields, etc. We're not going to focus on those, but we will be talking a little bit about how this might enterprise impact enterprise value, sorry. And how you can build a business case around this. So enterprise value is basically the multiplication of of the profit you make as an organization times the valuation multiple that you can yield in the market, whether you're privately owned or publicly owned. Let's break down profit as a first one. So profit is obviously sales and your margin percentage that you can drive for that. So the business cases that we are making these days around autonomous innovation and building new enterprise capabilities around that, the business cases that go to CFOs, they look at both of these things. So the expectation is that they lead to an increase in company growth rate and overall top line by bringing to market better products, sometimes more products, but especially also faster to market and having a clear impact on the net new sales of new products being launched, and then secondly a clear expectation. And obviously there's more certainty around the cost side of things, but the margin improvement through roughly a 50% cost saving throughout the product life cycle of where the autonomous engine is being applied to. And as such, we're also in some of these business cases, disrupting our own business. As traditionally an innovation consultancy, in almost any business case there is a line item that would point towards a decreased spend on innovation consultancies or agencies in general. So definitely a very interesting point in time in terms of disrupting our own business and our own industry as well. And then the second lever next to profit in terms of enterprise value is the valuation multiples. So there's obviously always an industry benchmark. I think we can make a general expectation that some industry multiples might move a bit closer to tech multiples as industries overall become more autonomous. But that's not something you can just influence individually, but the equity story that you bring to the market and that CEOs, CFOs bring into their earnings calls, they do influence the, the, the valuation multiple. And I think it's clear to, to state that markets will value AI native companies with a bigger multiple in the same way that they have valued digitally native companies with a bigger multiple. And we also see a clear desire desire with the companies that we work with to bring this story outwards on the new autonomous engines they're building, etc for, for this very reason. So if we're talking about building a business case around autonomous innovation, think about the increase it will have on the top line and the growth rate on the margin improvement in terms of the cost saving, which is also obviously a time saving throughout the product life cycle. And then the equity story that you can bring of transforming towards an AI native company. So that has more of an executive view on envisioning the future and how to look at that more from a business case and market valuation perspective. So after we've envisioned the future and built out a portfolio of initiatives that we want to look at across those four different boxes, there is a time to actually also build, build something. And I, I just want to take you a little bit. Yeah. In into that. And on this we're, we're planning another webinar where we're looking at it at this through a more technical lens on how we actually build these things and what more of the technicalities and technologies are underpinning that. But today we're keeping it at the executive blueprint level of how this looks like. So if we think about building, and this might be an over that you've seen before if you've participated in other webinars. But this is roughly the conceptual model where we start from. And obviously this varies company to company, industry to industry. But in rough terms it starts from an insight engine that takes in social listening data, new patterns, scientific breakthroughs, sales data, proprietary data pools, any data that you think would be relevant for your new product development engine that feeds into a concept engine which is basically generating new product concept, new service concepts, which is typically trained on, in, in if that's in the consumer space, on the brand identities, the demand spaces, et cetera. But We've also done this for example in the construction space or more industrial B2B environments as well. Those concepts and ideas then feed into what's called a simulation engine where we would simulate desirability, for example, by testing it against synthetic customer panels. As you might know is that LLMs are more and more scientifically proven to be able to represent consumer and customer audiences and and demographics. But we're also looking at feasibility simulation and viability simulation to kind of all look at that at the same time. And then there's a build engine that would actually build a first prototype of that product with a first design, for example, but also look at generative coding and with the likes of Devin that we talked about earlier, obviously the possibilities are only expanding in that space as well. And now last but not least, there's a scaling engine that could look at more autonomous ways of launching products, especially in a D2C type of environment. And obviously autonomous products in themselves, especially in the digital space. Self evolving websites, apps and and digital products in general, as talked about earlier. So that's love, roughly the architectural model of what we talk about when talking about an autonomous engine to just play that out on a high level on two quick examples. So very much an edge case is she in which launches no less than a thousand products every day, taking in new trends on social listening, generating concepts, basically not simulating them but validating them directly into their e commerce webshop. So they have a model of autonomously launching and autonomously validating them. And then the ones that are successful they would start reducing and distributing and they've shortened the time from a concept being generated to full scale production and distribution to just three days, which is obviously the true unlock of their model. There's a lot of things to say about chain. So definitely not advocating to copy this model, but just to play out how this works within this architectural overview. And then I want to take you to a couple of screens and a quick demo of some of the engines that we're building with clients. Obviously all of them are custom and proprietary to clients, so I can't show you the actual ones in use within companies. But if we just look at a quick demo that would look at generating new products in the food and beverage space and specifically in snacking, taking in social listening data, the model being trained on brand identities, demand spaces, etc. Simulating desirability against those synthetic consumer panels and generating first designs. As mentioned, we've built a couple of those and here you see that at work. So you saw the Insights engine, the concept engine, and the validation engine. So basically you brief the engine in terms of, hey, I'm looking to develop a new product for this Persona, for this demand space in this occasion, for this brand. And it would start to generate ideas in just a couple of seconds. And here you see a couple of examples in the snacking space and you can see the concept cards which would describe what the consumer problem is. It solves the concept description and early considerations around feasibility and viability. And that's especially something that we're developing out in more detail and where we see a lot of the value in the consumer goods space. And now what we've built as well is what's called a concept stretcher, which would basically take this concept and create variations on that across a number of dimensions of your choosing, like the pack design, the product, the ingredients, the form factor, etc. And then you would click next and then it would start generating concepts against those as well. So it becomes a dynamic copilot on, on the go. So we have early proof showing that that it works. So the, the concepts that are being generated, they hold their ground against traditionally generated product concepts and score well in consumer testing. We've seen this across categories from food and beverage to fashion and other consumer industries, as well as, as I mentioned, in the B2B and industrial spaces. In terms of the underlying infrastructure and asset, we'll have a separate webinar diving deep into all of the technical stuff behind it. But just wanted to explain to you that these engines work with different agents that work in the back end. For example, a desirability agent, viability agents, sustainability agents, and a feasibility agent that talk to each other when a concept is being generated, that critique that and concepts and products get iterated based on that. And we're doing this both as for physical products, but also for digital products, as it also works in terms of UX and ui. Our point of view is that the most successful engines integrate multiple functions. And this is also where we start talking more about the operating model around it, which I'll get to in a minute as well. So we're not advocating or basically advocating against just building an engine for one function and doing one step into the process better. Basically those productivity upgrades, we think that the unlock is looking at it holistically and doing all of those things at the same time. When you generate new product ideas, not just to look at the consumer desirability, but immediately check that against your manufacturing capabilities, the sourcing of materials, etc. Etc. And yeah, this is Just one example of how the two play hand in hand. So for some of them we have a formulation generator which basically maps both the desirability simulation as well as the ingredients that go into that still within the food and beverage space here, and that looks at autonomous formulation suggestions, basically maps out ten thousands of possible formulations for the new product and then test those against the digital twin of consumers and gives you a ranking of the best ingredient combination. This slide shows you way too much detail, but obviously we will be sharing the slides afterwards as well, so you can have the time to read through if you're interested. But this is basically a breakdown. Breakdown, sorry, of all the different engines from the insights engine to a scaling engine. The functionalities that go within that, but also the data that you can plug into that so it becomes a more dynamic whole. So that's about building software engines if you want. And then the last part that I just want to quickly touch upon is transform how to transform your organization. And I think our primary insight here, and also learning is that we see that the key to success actually won't be the technology that we just talked about, but the design of the new operating model to unlock its real value. And that's where the hard work is going to be. The technology is actually quite advanced and obviously you need to build that so it becomes an autonomous engine for your organization. But if you were to want to go to an Insight to product launch in just five days to just build on that example, that won't come from just building the technology, you will need to transform the operating model around it in terms of your production capabilities, go to market distribution capabilities to make that possible. And the bottleneck won't be the technology, but will rather be the human and organizational transformation around it. Because as if I've not stated that clearly yet, we do advocate to always keep a layer of human curation above all of this one example to just illustrate that in our minds the mode won't be in the technology and actually not even in the data. You could argue, maybe barring some exceptions, and a noticeable example of that is Last year in March 2023, Bloomberg spent over $10 million on training its own proprietary Bloomberg GPT, trained on all of its proprietary data with the ambition of building a superior LLM specifically for financial analysis, and only to find out that the next version of GPT, which came out a couple of months after that, was substantially better at financial text analytics than the proprietary trained Bloomberg GPT. So basically it's almost impossible for any company to out compete on the pure LLM AI model or foundational model technology, even with proprietary data. Because the latest technology in the state of the art like, like GPT or like Gemini Cloud, etc. Is basically available to billions of people around the world for only $20 per per month. So it's not the technology, obviously you need to build the system, but it's going to be in the operating model. And so the mindset we would invite you to adopt is to as an organization, think about becoming an autonomous innovation engine. The technology as well as the operating model, rather than just building the software. When we talk about transforming the system, and I hinted to this earlier as well, if you look at, let's say, the levels of altitude in an organization, from an individual or team level, a function level, insights, innovation, R&D, HR etc, to the whole company system, and then the ecosystem where the company operates in the suppliers, clients, partners, etc. Most of today's AI transformation roadmaps are very much focused on transforming the existing functions because use cases are sourced bottom up. The existing insights function has a use case to leverage AI for better insights analysis. R and D, has our use cases within R and D, et cetera, et cetera, basically focused on that productivity upgrade box that we talked about earlier, focused on building AI powered solutions that do the current faster and cheaper. And in line with what I talked about before, where we see the true potential is looking at it holistically and building a new AI native operating model that will be a real industry game changer. Again, like this is a simple mental model, but I hope it can help you when whenever you see discussions internally that you can anchor them. Well, actually, at what level are we talking about? Is this just an individual level? People writing emails better or faster? Are we talking about optimizing within a function? Whereas maybe our question should be is that actually a function or a step in the process that we still need? And how much of the discussion is actually talking about looking at it holistically across the company. So in essence for us, autonomous innovation is less just about the technology, but it's about building a new enterprise capability. This is a model that we've been working with with a couple of companies that looks at the strategy you set around that and then looking at talent and culture, the tech stack, processes, structures and governance, rewards and metrics, ultimately leading to better products, faster time to market, cost efficiencies and a unique competitive advantage we're really strong believers of. And you need to look at all of these things. For example, with a lot of Transformation initiatives, incentives just to pick out one, are typically averse to what you want to achieve. And we need to look at this holistically, not just looking at the model. I said we'll share the slides so you can dive a bit deeper into all of this. So I'm almost wrapping up, but just wanted to look at how does this all come together and what in my mind is the current executive blueprint for autonomous innovation to just go through it one more time. So for us, it starts with envisioning the future rather than starting from the current state. What is that increasingly autonomous world and how will we win in that? Think about that two by two matrix and that portfolio approach. Setting the vision, setting the strategy, but also getting leadership on board around that vision and how we're going to win in that future world. Thinking about transforming the system across people, processes, data, technology, and looking at that holistically rather than function per function. And then thirdly, building a new engine, an autonomous engine for always on innovation and growth that goes from imagining your future offerings to making them and successfully selling and scaling them with a layer of sensing and validating both synthetically and in the real world continuously. So if there's one thing that I would love to invite you for for in this webinar, it's to look at those three things, envisioning the future, building a new engine and transforming the operating model or the system. And before I conclude and we go to a bit of Q and A to wrap it all up, there's one more thing. You might have received a message or an email already, but we're super happy to announce our second edition of the Autonomous Innovation Summit. So the last one was early December. The new one will be June 5th and 6th of this year. So in roughly two and a half months from now, we will again be inviting you for two days, 40 plus speakers, 40 plus sessions across two different stages. We already have a couple of super interesting speakers from the likes of Google. We have two speakers that are building fully autonomous ventures. One person that's on on a path towards building a one person unicorn company leveraged by AI. So a lot of interesting stories. We have people from Harvard, mit, et cetera. So we hope to all see you there again if you register before the 1st of April. April we're giving away the super early bird tickets for free and you can find them on autonomous innovations summit.com and on that note, I hope I've been able to give you a flavor of what an executive blueprint for autonomous innovation is, how to envision that future of a more autonomous world, how to build an engine for that, and how to think about transforming your organization around that holistically. And on that note, I want to wrap it up and open us up for questions.

I can potentially assist with guiding through the questions. We can start with a simple one actually a funny one after B2C and B2B. Is there an AI to AI market?

Well that's. That's a super interesting one. And my. My answer would be. Would definitely be yes and was actually just maybe an anecdotal example that I saw about an hour ago, which is somebody that basically built a deep fake of himself doing an actual job interview with a human. And it was not distinguishable. So you could immediately see that on the other hand there would actually be a deep fake of the recruiter as well and it would actually be virtual avatars of both of them talking to one another in a first stage of a job interview process. That's probably a ludicrous example and actually I don't think that's really an added value at all. But we will be seeing AIs talking to AIs in various occasions. Yeah, absolutely. Think about customer service for example. We will for sure have personal AI assistants to speak to the AI customer service chatbots of enterprises.

Maybe a more serious question next on what's your view on the EU's new AI act and is the impact on the development and competitiveness in the world?

Yeah, that's a good one. And obviously for those that aren't aware I'm based in Europe although I'm very often in the US as well. So I can definitely see the two differences and I guess if we relate it back to what happened with Internet and digital, I guess most of those solutions and the whole economy was based on was built out of the US So the world's most valuable companies around that are in the US and Europe was the first to regulate them with the likes of GDPR which I think were like lighthouse regulations which I would in general advocate for. And I guess we're now seeing exactly the same with AI is that a lot of those obviously not exclusively but are built in in the US and that Europe is the first to actually enact a law. It's not in place yet. Right. So there is a discussion draft and it will probably get voted in the fall of. Of this year. Was also they started writing it by the way before Gen AI became a thing. So a lot of the the regulations in there around are around protecting society. That for example the Police can't use AI to basically monitor us everywhere and with everything that we do. So it's not specifically for Genai that maybe as a first and I think it's a good attempt at classifying to a risk level of each of the applications and then tying regulations to that. I don't think that's necessarily what's holding back competitiveness. I think we are definitely or Europe is not competitive in AI, but that's more because of the historical lack of actually not having one in the digital space yet, which now leads to not winning in the AI space either.

Okay, maybe transitioning into a question more corporate question on one from David. By the way, we're selecting some of the questions based on the highest upfront votes. It'll be, according to a question from David, is it will be difficult to move from standard product development for many companies. How do you suggest to bring companies into the breakthrough innovation space, especially larger companies?

Sorry, I lost you for a second, could you repeat the question?

So the, the question was it will be difficult to move from standard product development for many companies. Therefore, how would you suggest to bring companies into breakthrough innovation into the breakthrough innovation space, especially larger companies.

Yeah, that's a good one. So if we think about those four boxes on the matrix, there's also an interesting overlay to think about how you're going to execute against them organizationally. So obviously everything in the box of productivity upgrades you would own by existing line management, existing functions, etc. If we think about breakthrough innovation, that's likely best served as a separate initiative first. Obviously you need to envision what that is, but in similar ways that for example, a car company like Volvo would launch their EV line as a separate brand with Polestar, for example, as a sister company and Renault is doing similar things. It's always difficult to disrupt your business within the existing. So breakthrough innovation is likely best served as an entirely new venture that then later on you would integrate and then the other two ones. So product innovation is likely best served as a new product line or a new business line that still anchors somewhere in the existing organization. And then system innovation would need to sit at C level across all the functions. So that would best be owned by a chief Transformation Officer or someone in that role or capacity. So that's kind of roughly how I look at how to organizationally execute against those four.

Can have you to pick another question or otherwise I'll shoot the final question at you.

Yeah, go ahead.

So it's more of a philosophical question. It's shared by Carlos the data that is feeding the large language models are human generated today. But as we move to a world where data might be produced by other large language models, do we see some sort of a breakthrough in creativity decelerating due to the majority of people not generating new content, but actually consuming or reconsuming AI generated content? So curious to get.

Yeah, that's a really good question and it's quite an intense debate whether that's going to be the case. So I don't think there's a conclusive answer to that. So there are studies that I've seen that show that actually synthetically generated data can also be used to train models on and can equally lead to divergence. We've definitely seen studies that if you prompt LLMs well, that they can lead to at least equally creative ideas. They don't do by default. So you do need to prompt them for it then if you compare it to a group of humans. So I would say the jury is out on that one.

Okay, we, we have maybe time for one last question then. Question on how are you thinking about the role of proprietary consumer insights for businesses and the impact on innovation engines?

The role of proprietary consumer insights. Right.

Yes.

I'm trying to think what's meant by that in terms of. There's a couple of perspectives. So obviously there's a bulk like there's a whole load of consumer insights that any consumer goods company would have that isn't fully leveraged. So I think there's definitely a potential to all bring that into an engine so that we avoid any duplication of research. I would want to keep in mind the example of Bloomberg that actually I think it's like we can easily think that actually training a model based on our proprietary data will lead to a better model. But actually there's little proof of that at the moment, as the example of Bloomberg has shown. So I wouldn't overplay the value of data to generate a better model. Like it might be relevant for other things, but maybe not for a better model.

Okay, I don't think we have time left for a final call, so I would just love to thank you for the presentation. I've seen a lot of very positive responses and looking forward to follow more.

All right, thank you so much everyone for being here. Would love to see you in the next one. Thank you so much. Bye everyone.