Webinar: Circular Business Design

All right, well, why don't we go ahead and get started. We're excited to have you all joining us. We are ator. I'm joined by etor. My name is Monty. We're both from Board of Innovation, where we seek to, where we're a global innovation firm seeking to build tomorrow's products and services and businesses and create them. Today, we're really, really excited to have you all joining us to talk about the circular economy and sustainability and how we can make progress in this transition by leveraging AI to get started, I'm going to throw a question out and I'm going to ask that you drop in the chat who you think this person is. Love to hear any guesses or people who know this person. Henry Ford. It's a good guess. Not exactly who it is, but someone, someone similar. The guy from gm. All right, Nikki. Oh, I seen. Okay, Nikki, good to see you on here. Indeed. Yes. The guy from GM is the closest guest. This is Alfred Elon Musk. No, maybe like the Elon Musk, similar to the Elon Musk of like a hundred years ago, but this is Alfred Sloan. And Alfred Sloan, he used to run GM General Motors back in the day. So from the 20s and the 30s and even into the, into the 40s, he was responsible for, for General Motors. And he's also responsible in a large part for how organizations are structured and organized today to deliver products and services. And he's one of the things maybe he's most famous for is he helped to popularize this concept of planned obsolescence. Right. There was others involved from like the light bulb cartel in the 20s as well. But Alfred Sloan really took this idea of building a practice of designing products to break down early because. Become obsolete sooner so that consumers would like to buy more. He popularized that idea, largely popularized that idea, and is responsible for a lot of, I would say maybe the, the consumerism and the, and the rapid product development that we, that we face today, which, you know, maybe was less of an issue or a smaller issue back a hundred years ago. But I think we can all relate to feeling the pain and thinking, okay, I'm. There's got to be a better way. There's a different approach that we can take to develop products and services that don't serve the last century, but serve the next century. And that's what we're hoping to cover with you all today is how do we take this concept of circular economy and designing for sustainability and then overlay that with the tool set and the power that we're starting to experience, coming from AI powered innovation to really create products and services that are more meaningful, perhaps more resilient, but, but better serve the planet and humanity to kind of get some big ideas out there. So how we'll structure today is I'll introduce just the kind of the basics of circular economy so that we're aligned and speaking the same language. Then we'll talk to you a little bit about this concept of sustainable innovation. And then ETOR is gonna really drive us home around how do we take a practical approach to leveraging generative AI to develop products for circularity and sustainability so that we can not just build more, but we can build better products. So just to kind of lay the groundwork for the circular economy. So circular economies, or at least a circular business model, is really the way in which that organizations create and deliver value to its broader range of stakeholders while minimizing the ecological and social costs. So in a traditional kind of product development perspective, you might see it as very linear, right? We take something from the earth, either through mining or farming or what have you, we make something from that, it's then used and then it's sent to a landfill or, or, or, or elsewhere, right? It becomes, it becomes waste. And so it, not only does it have a negative impact on, on planet, but it also has, I think, a much less of a consumer oriented approach in many ways and can kind of reduce the value that we get out of those products and services that we're building, right? So with a circular economy perspective, we think that we can do better and we should do better in making this less of a linear process and making it more circular. So through recycling, reuse, repair, return, we think there's approaches, or there are approaches to unlocking more value across the value chain to better serve our stakeholders or shareholders, even to partner with our ecosystem, other ecosystem players, to find other ways to unlock value that serves our businesses and our consumers, but also has less of a negative impact on the planet. And we believe that achieving circularity and taking the principles of circularity and applying them more rigorously and more pointedly to our product development and as innovators is essential for us to achieve sustainable growth. So ETOR is going to now walk us through a bit about sustainable growth and then we're going to get into some more practicalities.

So just taking a little bit of a step back here and to look a little bit in theory. So the first thing that we need to understand is that many of our socioeconomic systems, they scale in a way in A matter that is explanation, right? Our cities, our economies, other social systems, and we can kind of feel that growth and that acceleration in our bones. We can cannot stop, but constantly talk about how things are changing and accelerating even faster. But this exponential growth in the theory kind of presents a paradox like, can it be infinite but in a finite time? And then if we think about it, actually, it can't. So this paradox is kind of resolved by reaching a certain point. So the system reaches a certain point in which it stagnates, so it stops growing naturally, and then it eventually starts to collapse. Right? But this rationale on its own would be sort of like the classic and very old Methuselah theory, right? Which is the theory that used to say that, you know, populations cannot grow forever, we cannot feed populations forever with limited resources, right? We will collapse. Something is not complete in this theory, right? So what is not complete is that basically innovation. What solves this kind of, this issue, this paradox here is that innovation kind of resets the parameters of a system. So before we reach a certain stage of stagnation and collapse, there's a certain paradigm shift in the system that allows us to keep pushing it, to keep growing. But our socioeconomic systems behave in a certain way that each one of these cycles, they actually need to happen in an accelerating pace. And I will actually pass back to Monte because AI is really fitting into this story of this accelerating base.

Yeah, exactly. And I think if anybody here has been on LinkedIn or on a computer in the last year, Genai has been dominating the ethos of those platforms in thinking about productivity and product development and innovation. It's the latest, latest thing. And I think there's a lot of, you know, maybe theater that happens around, around it, but the reality is that it is kind of starting the. We're starting this journey of our lifetime around channeling innovation tools to create new products and services. And Gen AI is going to be. And AI more broadly is going to be a primary focal point or is a primary focal point of that, right? So Gen AI in particular, it has the potential to drastically increase our productivity and do so in a faster way than other kind of major technology breakthroughs of the last couple hundred years, right? So where electrification maybe took 40 years to start to truly have an impact on productivity. And as a result, innovation and computers took maybe 20 years to get to that point. Gen AI is starting to show impact in less than three years. And so the question becomes, as Etor was mentioning, this need for like, tighter innovation cycles. The question becomes how do we harness that power from Gen AI that allows us to tighten those circles, but do so in a way that's responsible and that helps us achieve more sustainable growth. So AI is the ultimate kind of accelerator to those cycles, but it also has potentially negative connotations or negative impact if we're not careful. And that's what we're going to talk about a little bit here in a moment because it's not so much that AI is going to change the way that we operate, change the way that we develop products, but it actually already is. I'm not going to go through all of these examples, but I will call out a couple. So if you're kind of interested and you're joining this sustainability and circularity webinar, then chances are you've heard of Shine and you're probably, or Sheen and you're probably not the most, the biggest fan of them as a, as a company from a like a fast fashion perspective, which generates enormous amounts of waste. But one thing that is interesting about what they're doing is how they're kind of informing that level of product development, which is by leveraging kind of this idea of an autonomous innovation engine to leverage social listening and market demand and then rapidly translating that market demand into new products and services. But the, the, the, and so that's kind of creating like enormous amounts of, of waste. But it's also opening up this paradigm of like, okay, how can we move more quickly and what are maybe some of the benefits of being able to move more quickly that we'll talk about in a second. And then we see other organizations, whether it's in technology, with GitHub, loving it, leveraging AI to, to write code or, or in healthcare, all the product discoveries and innovations happening there, we're seeing very like many positive benefits of leveraging it in our innovation as well. But the challenge is that despite its great capabilities to launch new products, whether it's in fashion or healthcare or technology, AI is kind of dumb, right? It's totally a garbage in, garbage out scenario. And so without the right parameters, without the right data and asks being made of it, it can create terrible things. So coming back to the Shine example we discussed Shine, how it launches these products rapidly, thousands of products a day and week based on market learning and kind of what's in demand, but without the kind of the key clear limits that it's placing on that system, it can create really terrible things like the swastika pendant, which I think we can all agree, and I'm sure Shine would agree is completely out of, out of scope for what is right and what should be developed. Leveraging these, these tools. But going back to a tour's point around kind of this growth conversation of how do we accelerate our growth in a sustainable way. Gen AI, AI more broadly has the capacity to accelerate those, those cycles. And we need to recognize though that if we don't set the right parameters then we're not going to be able to accelerate those cycles in the right direction and we'll end up achieving maybe growth and productivity gains, but in the wrong direction and with the wrong impact. There may be a case, and you know this is that there where there's worse scenarios than the Shine case. But wanted to highlight it as kind of the clear call to action for this group and fellow innovators to say we need to be consciously setting the parameters of our, of our product development engines to build what we want and what we, what is aligned to our ethical standards and our, and our circularity or sustainability goals to be successful. Right. So shifting this idea of leveraging AI to do everything it can do to being responsible and harmless, which you know, is a conversation of around alignment that companies like Microsoft and OpenAI and Anthropic are having. But it's up to us as innovators to also lead that goal, lead that charge and start to set the products and the parameters that we want to see in the world and related to our sustainability or circularity goals that we might have just one quick example of, maybe something that is responsible and harmless. Right? Relatively harmless. I think the effects of soda can be debated but wanted to share this example of Coca Cola who came up with this new product for a soda that tastes like the year 3000. Right. It's probably being fair to Coca Cola, maybe more of a marketing campaign and then than anything else, but I think we can agree that while this product maybe is responsible and harmless, it's maybe not what life should look like in year 3000. It's not how a soda should be delivered maybe either from a product standpoint or a packaging standpoint to consumers in year 3000. And for us it left us a little like kind of wanting more about what could a product really look like in year 3000 and what might the packaging look like and what might it enable from a circularity and sustainability standpoint. So we kind of at our kind of core, and I would say the purpose for this, for this topic today is that we believe that if we leverage kind of the bold thinking of what AI powered innovation can do, but set the right parameters and desires. With that tool set, we can not only build products that are responsible and harmless, but products that are actually good for the world and help us to achieve that sustainable growth that we're after. So I'm going to pause there and I'm just going to ask a question for the group and if you can drop in the chat, I'm just curious if there's any, if there's been any AI powered innovations kind of broadly speaking or maybe more specific to circularity that have fascinated you or that you found interesting that you can share with the, with the group. And feel free to just drop those in the, drop those in the chat. I'm actually going to stop sharing just for a minute so that I can see it. And as we go along too, I should note the Q and A function on the, on Goldcast allows us to free to add your questions there. So if those come up as we get into more of the practical examples, feel free to, feel free to drop those, those questions in the Q and A function. And we have one from Thomas, the Altana Atlas global supply chain mapping tool. Great, we'll have to, we'll have to take a look at that.

All right, feel free to kind of keep dropping your ideas in the chat as we go along, but right now we want to get a little more practical and discuss how we can use AI to design for sustainability more broadly and circularity more specifically. The first thing that we need to understand in this context is that creating good enough ideas is becoming fast and cheap with autonomous innovation and AI powered innovation more broadly. To illustrate this, we are partnering with a major global product company to develop an innovation copilot that is trained to generate, test and assess new products for many of their brands. What this graph shows you are the results of a task that we did with this concept. We gave people within the company 20 concepts for them to evaluate blindly because 10 of those concepts were AI generated and 10 of them were some of the best human generated concepts and ideas that they had recently. And the Y axis basically shows you how much the concept was liked or disliked versus the other concepts. The blue ones are the ones that were AI generated. And you can see that out of the top 10 ones, four of them were AI generated and the highest performing one was AI generated. And we need to remember that this is the worst AI will ever be. So after only four weeks of developing this autonomous and this custom AI engine, were able to come close to the performance of their best concepts that they had spent hundreds of thousands of Dollars and hours to actually develop them. Right. So this idea of making good enough ideas very fast and very cheaply is essential to understand where we're going. Because this efficiency should give us more time and resources with which we can do better and not only more. Right. So if we go to our circles that Monte was showing before, you know, we can do everything AI can do and kind of explore that. We can have it responsible and harmless, but we believe that we should use it and leverage it, this efficiency gains to be able to make what life needs next. So in general, like when we think about benefits of AI for innovators, we think that it can be like faster, right? Faster go to market or we can create more ideas and more options to bring to market. And Shane example is a very good kind of good example of this more. And then we have a question mark of what kind of more is that? And also increase the hit rate of that. Right. Because you can simulate that and quickly iterate on that. But of course, as we've been saying, with these efficiency gains we can do better because you're able to iterate on those and towards making what life needs next. So we're going to take another quick break here and this time you can take your smartphone and point your camera towards the QR code or go to slido.com and put their code there. To answer the question here, what are the organization's biggest challenges or the challenge that you face with your clients when you're trying to innovate for circularity? Right. So start dropping in some of your challenges that you face and they're going to start popping up here on the screen for us to have a quick conversation around them. I see some people typing and the answers are going to come in a second. Revenues. Yes. Being sure that we go towards that. Not only sustainable, but actually like viable business model around a sustainable product is one of the most challenging elements that we have. Having strong and vertical competencies like the competences around that space. That's also a very interesting one. Yeah. Finding the, finding stakeholders cost assessment associated with the operational model. Very, very interesting. Moving too slow. Yeah, that's a good one. Very, very interesting. Yes. Ecosystem approach. That's also another one that is, that's also very cool. Yep. Top down, top down support. Yeah. Sometimes you're just like, you never prioritize that in the company. Right. It's like we all want to do that but then, yeah, basically you don't have enough top down support. Keep them coming. But also like, I'm going to change the slides and you might miss that on your screen, but then keep them coming on the chat if you lose the access to slido. Because moving forward here into some of those challenges, some of them we can at least start solving them on the early stages of innovation. Right? Because what we know is that 80% of a product's environmental impact is influenced by the decisions made at the design stage. Okay. So if we can make sure that on these early stages, we can already point sort of this product development path towards a more impactful one, we can be sure that like later on this product is actually going to make deliver the impact that we want out of that. Right. Instead of just being an afterthought. But currently, and I think that's related to a lot of the challenges that were put on the chat. So for instance, like not top down support or like the time it takes and like trying to fit all of these different elements together. Oh yeah, the stat comes from Ellen MacArthur foundation, already seen the chat, but it's the Ellen MacArthur foundation which is the greatest, one of the greatest references for circular economy. But I can share that later on. But currently in the first stages of the innovation process, as some of you pointed out, even good enough ideas take time and are expensive to develop. Because of this lack of support, this lack of time, we ended up dropping the sustainability lens to try to prioritize the viability, the feasibility and desirability. But as we saw before with the autonomous engine as a first example, we see that this process is changing quite fast. Right. So let's take a moment to pause here and try to reflect. Okay? So if we're changing that, we're thinking about that this process is changing kind of what it could look like in this future and how can we harness this new process to actually deliver what life needs next? So this AI powered innovation will get us along further in the validation journey in last time, right. Every time we talk about innovation, we always talk about like this early stage, this, you know, front end where you kind of develop the idea, you have empathy to it and then later on you start validating that. Right? But with an AI powered innovation, you can really get us further along this validation journey in much less time. Because we see basically now three major steps in this AI environment, right? So the first one is we see that's like about training. It's then developing your ideas and then iterating so getting more into the details. The first one is really about training your models and generating your hypothesis around what you want to develop around the scope of the project or the scope of your company, then there is a moment of the exponential concept development. And that's where AI really thrives. This capacity to generate hundreds and hundreds of ideas very, very fast and very cheaply. And then finally we have this moment of iteration that you can take much longer because these first two steps are so quick, fast. You can kind of take longer on this iteration path using not only human but also synthetic tests to go throughout this journey of validating these ideas. So in the end you have something that is both desirable, viable, feasible, and most importantly, sustainable. So if we can incorporate key data that promotes more sustainable outputs as we train our models and our specific for our context or our product specific models, we can then leverage this data later on the later stages to build ideas around this data, but also iterating on the ideas that we have or other ideas that we already have in the company. So let's take another step into kind of looking under the hood of what this autonomous or AI powered innovation engine can look like. So imagine that you have this engine that is formed by AI agents, okay? And these AI agents, they are tasked to generate permutations of a new product, okay? And they are, each one of them is optimized across a different lens, right? The four lenses that we were discussing before. So each agent, imagine that each one of them are a separate GPT. So basically a different chat, a different kind of entity, and each one of them has a different goal, right? So we have the viability agents that is always like asking like, can we please make sure that the time we build makes money, but the thing we build makes money Desirability agents, like, oh, the only thing I care is about creating something customers love. And then you have the sustainability agent fed with the data that we were discussing before, trying to build something that has positive impact on the environment and society and finding the feasibility agents that really want to get practical. All right, so imagine that you have for these agents, agents and they are talking to each other and trying to create sort of a permutation of new products for us. And we're going to try to illustrate here for you. This would be sort of like for healthcare service. Again, this is an illustration. It's not like literally how things go. But just so you can understand this rationale of the chat in the background. So let's say you're creating this healthcare service which is around kind of access to clinicians. So the first viability agent says, oh, let's make sure we don't give anybody access to real clinicians. Though that would drive up costs, right? Because that's what it's optimized for, like reducing costs, generating more revenue. But then you have in the desirability agent comes in and saying like oh, we had this premium perception in the beginning of our conversation before. That will kill this premium perception. So we need an alternative proposition. The sustainability agent can come in and say like oh, if they were accessible on demand and they don't need to kind of move around these clinicians, this would reduce our emissions by reducing our operational footprint. Whereas the desirability agents can then come in and say like oh, being remote and on demand could actually be our new differentiator. The feasibility just comes in and say we just need to make sure there is a service partner that could provide that on demand. But honed on, I found three of them because these agents can be connected to the Internet as GPT4, Bing AI already are. And then finally you have a viability agent again coming in and say like, okay, so I simulate a competitive bid between them. We can reduce our costs overall by 10% and maintain our value perceptions. Right. Again, this is just an illustration. It's not literally how things go, but this dynamic interplay between these different agencies, agents. You can improve and assess a concept to either come to a consensus or come up with different options that humans can consider and put on the table. And because everything, all of this is happening on the backstage and doing very very fast, happening very, very fast, you can have more time to kind of actually reflect on that. So we're talking about this autonomous innovation engine that's kind of this more complicated projects that we are kind of doing and we're kind of building this custom engines. But these engines are built on top of the LL systems like GPT and we can actually use them to experiment and generate something in our day to day, even if we don't have access to a more sophisticated engine. I'm sure most of you have experimented around with GPT already to generate ideas. And of course you can use the, you know, the model to identify or define your Persona as kind of understanding from a desirability standpoint who are we trying to deliver products and services for? Then based on that, you can use the chat to kind of start generating the concepts, like start breaking down the different elements of each one of the concepts. If we're talking about a food product, which is the example here, what does the packaging look like, what is the ingredients like, and so on and so forth, what's the value proposition? What's the experience that the consumers are going through and you very quickly can start to generate those concepts using a GPT like model focused on our conversation here today. And we're going to break down this a little bit later on. You can start to break down the different components of your product and start iterating on different alternatives that you could have for that. So for instance, in our food product example here, we know that a food product largely contains its ingredients and its packaging. Of course, transportation could also be an element here and how it's sold and so on and so forth. But you just break it down into the different components that you want to look into and you start to iterate on it to provide alternatives that could be more sustainable. And then finally, based on this iteration, you can then start to design the packaging, iterate on the concept again based on desirability, because some of the choices from the sustainability might reduce a little bit of the desirability. But then the message here, like to really understand, and I think it was clear on the illustration from the healthcare one as well, it's basically really a continuous iteration exercise with these four lenses doing it very, very fast, making sure that one of those lenses is actually sustainability. But let's look into the sustainability one, which is why we're kind of the focus of our conversation here today. So first of all, it's very important to understand that you shouldn't expect models like a GPT to do all its tasks in one prompt, right? So like, oh, give me the perfect sustainable product that's also going to be viable and desirable and everything. It's going to be high level. It's going to be like, it's not really going to work. So what you need to do is translate any methodological process that you agree with and that you believe in into this conversation and walk the model through it. Remember that this is kind of a copilot and you're going to be using it as kind of something that you're going to be taking along with you. You, right? So for instance, in this example and this experiment that we showed you, the first thing that we did is that we incorporated the data that is relevant to the experiment. In this case here, we incorporated the known's public annual report. This was a public available information that was there by the way. This experiment was not created with the KNOWN was just a quick experiment that we did and we uploaded it to GPT4 and we just based on this information, like it's a long deck, probably like 50 pages long. You don't need to organize that Data. But basically you use the model to start to define what is going to be the evaluation criteria that you're going to be using to iterate and develop your ideas, right? So the model can go through your DAC or whatever document that you want to put. And by the way, very importantly, very important point, don't upload anything that is, that is sensitive to ChatGPT but like whatever you put there, it can go through it and you can start defining what is the criteria that you're going to be using to evaluate your idea. And then finally, as we discussed before, start to break down the different components of can be a product, can be a service, really can be anything like as we saw before in the healthcare product and the circular model, you can have different elements of a business model that you want to look into, like the logistics. But in the case that we are showing here is what's a food product. And more specifically here I broke down into the ingredient components. So specifically was a drinkable yogurt where you have of course as an ingredient, you have the milk and the yogurt, you have the fruit puree, you have the sweeteners, you have the probiotics and you can start sparring with this partner and saying like, oh, this is the criteria that I have, which one's better? Look online for it and so on and so forth, right? So you can, because you can generate those ideas very, very fast, then you have more time to actually start iterating on this, which is much faster as well than going after. And I think that's one of the challenges that someone mentioned in the chat before was having that knowledge from the verticals in the company, right? So if you can take that knowledge from the verticals in the company and kind of consolidate that into kind of long documents and can be very long documents by the way that you feed into the model, you can then start sparing with these people that have this siloed information and then develop your idea on that. So but the thing is, and taking a step back and kind of a, kind of a reality check here is that GPT is not specialized in that and the sustainability report that we have here can be still somewhat high level, right? So what's very important is that if you have better inputs and for instance can be kind of conversations with specialists in your company or other documents that you have and you found we can have more, much more effective results. So if we go back to our model that we were discussing before, it's all about really making sure that this training model in the beginning really is fed with the information that is relevant so you can develop and iterate it further along the journey. So taking a moment here to give you a very specific example of a company that we are partnering with. Right. So Footbearing AI is a company that they. It's very kind of interesting what they're doing, but basically they're digitizing products and ingredients into their molecular level and they're analyzing them chemically to create an AI model that can generate surprising matches on the formulation. Right. So we can really put like, oh, about what about this ingredient where it can match with this other one and it can do it like with much more options than a human mind can go. Right. So this is really about training a model around the specifics of ingredients on their molecular level. So what can you do with that? Right. So with that more robust data and again doesn't have to be this specific example, but whatever example and whatever information that you have, you can use that information to really develop and iterate on your ideas, which are the follow up steps. So if we look at the Ellen MacArthur Foundation Circular design opportunities for ingredient selection and sourcing, we know that a lot of the impact might not be this number. Like, you know, these numbers always change. But we know that for a food company the impact that it has, a lot of it comes from the ingredients that it source and select. Right. And in the framework from Ellen McCarthy foundation, we know that it needs to be diverse, it needs to have lower impact, it needs to be upcycled and it needs to be regeneratively produced. Right. Or at least a different combination of those ingredients. But all right, you identify those ingredients. So what do we do with it? And that's where we think that these new models and new ways of working can happen. A really big impact. So if we say if we take an example here, in another case, Unilever and WWE WWF work together on future 50 ingredients. Basically they identified ingredients that were more sustainable and healthier, but they were also unfamiliar to the general population. Okay, so and they challenged food pairing to understand like where and how can we use that those ingredients, those products. Right, you can use on the development stage. So what products could we be developed with these ingredients in order to make them more familiar and more accepted? But also its rating on them. So what existing recipes, they would be a perfect fit? Or can we replace certain ingredient on my formulation and kind of keeping the acceptability of the existing product? Basically, for instance here we can take something that is like a pumpkin leaf which might not be an obvious ingredient but might check a lot of the boxes in terms of sustainability and you can take that and ask what is the recipe that has better fit for that. This is a very specific model but I think the message here is basically making sure that you have that information and you can fit that to the model that you're working in. So we can quickly iterate on these early stages of innovation in order to have the impact from this earlier stage. So just to wrap up a little bit and we have to take some time for kind of having a conversation and take some of your Q and A just to wrap up on some of the conclusions that we have here. So first of all AI will allow us to have a tremendous efficiency gains on this innovation process. Right. And we need to make sure that we not only create good enough ideas but we can actually do better with those efficiency gains. The second one is that really is very, very simple to start experimenting with AI for impact because you have those efficiency gains. You can kind of take this early stage of innovation to take some of this efficiency gains to experiment with GPT in order to have more impact. And finally after you start experimenting you quickly going to see where it should improve and then make sure you can start improving the model and your processes with data that you have. Because always very important to remember is that this will, this is the worst AI will ever be. So we need to make sure that we keep developing and better training the models around sustainable and circular data that we can, that we can leverage. So yeah, let's have, let's have the conversation. And Monty, I don't know if you want to walk us.

Yeah, I think so. I was trying to go through some of the, the Q A asynchronously with you but I think we have so feel free to it or as you as you go through the questions as well to add on, add on those. And then Irwan I you had an interesting question around whether agencies should now focus on post growth innovation rather than being stuck in a sustainable growth mindset that isn't maybe that is incompatible with planetary boundaries. And I think for me at least in my perspective and then tor feel free to add if you have if you wish. I think good, good innovation whether it's inside a corporate or a startup or or what have you or in entrepreneurship it leverages a combination of consumer demands, market trends and, and alignment to a, to a business strategy. And so from a kind of as we think about where we should be focusing as, as agencies I think picking you know the direction in Terms of where if we're running out of resources and we need to have this kind.

Of.

Kind of post growth mindset or relative to the available resources and innovate against finite resources, that to me is a, as a. Is a strong or compelling case for, for, for leveraging market evidence to, to make decisions in that way. Right. So, so considering that the innovations that we're creating maybe are not sustainable, but then figuring out how do we so, so shifting that mindset and building with finite resources in mind, with sustainability and circularity in mind and then generating resources from our products from that is like the foundation is the kind of like the thesis that, that a Tor and I are trying to, to allude to. But that requires a shift right in how we predict or how we manage financial viability and, and all of that. So it's a, it's like a complicated if, if I'm interpreting your question correctly, it's, I think it's a complicated subject of how you like create that shift. But I do kind of agree with you if, if you are implying this that a shift needs to happen to create more sustainable oriented products and services that are more resilient, that can be better traced and tracked and cycled up the value chain.

Yeah, I think basically that question, like if we take a step back from your question and actually ask can we keep growing forever or are we doomed to kind of that kind of collapse and the stop growing and then collapsing. I think that's a very, very complicated economic and socioeconomic question that a lot of economists are debating on who have amazing conversations going on around this. Honestly, it's a bit beyond the scope of what we can necessarily answer here. Are we actually going to be able to grow forever with this change of mindset that. Or are we just doomed to kind of have to change our paradigm and change our capitalist system? Like you can go on and on with this conversation. Very. Yeah, but I think what we are trying to do here is that like we know innovation is very powerful. We know innovation is an engine for this growth. And I think related to what Monty said, like can we change that on the early stages of innovation to make sure it has a lower lowering and even more positive impact. So I do agree with you that innovation agencies should focus to at least a new paradigm if it's post growth or a sustainable growth. It's very complex topic, but definitely needs to change, especially considering the efficiencies that we're going to have with this AI paradigm. There is another one that I think it's Interesting that someone asked about like innovative ideas. Yeah. How do you create autonomous innovation engines? And if that's the only way to kind of look into this, we don't believe this is the only way. We do believe that this will be a competitive differentiator for a lot of companies because if all the companies are using the same models, they will be stuck with somewhat similar answers. So fitting in your own proprietary data and you know, changing it slightly to your priorities will give you the competitive advantage. And then of course with this you then also, you know, allows you to put your own sustainable development goals into it and so on and so forth. So, but, but ChatGPT and other tools can be leveraged to start experimenting with this for sure. You can go very, you know, very far along the road with those roads with those tools. And I mean we have here like get in touch with us and then like if you want to have more conversation on that. But, but then at some point we do believe that those engines are, are important to create differentiation and specifically on our topic here today, and also making sure you put your own data, your own sustainability and circular data and goals so you can create ideas that are ready fit to that, to those goals.

Yeah. And then there's another question here from, from Jennifer. I was, I was typing, typing a response, but didn't, didn't wrap it up yet, Jennifer, but the, so how do you create the autonomous innovation engines? Is it with chat GPT4 or other tools can be used to create them? Yeah, I think it's, I think a Chat GPT is a, is a foundational tool, but depending on the specific elements of the engine that you're trying to, to build or test for other tools maybe are better, are better, better suited. Right. So if we think about an autonomous innovation kind of replacing our innovation process of how we go from generating insights to developing and testing and building concepts, to scaling, scaling those concepts, different tools are going to be needed to, to accomplish those different, different tasks. So in a, in a insight development standpoint it might be more like tools like outset for, for generating customer research or other tools for leveraging, like being AI for leveraging social listening. So it's a bit of a combination of different tools that will be required to create a autonomous innovation engine.

There is one interesting question from Chazelle Bruno. So make different agents talk to each other is very clever. Is it really possible to do with ChatGPT? Yes, it's going to be slower. It's funny to have slower in this context because it's already super Fast. But basically if you open four different conversations with the chat and you do it manually and you start answering to each one of them the answers from the other, you can basically emulate that exercise. And you can also do in one chat and you can ask the chat to create. Yeah. Imagine that you have this four like how a conversation would go that is more limited and it might get a little bit confused on that process. But like this process of creating four different chats and setting them up as different agents and then giving them different data points and then just moving on with copy paste as you feel doing. It's a way to kind of hack the system and create your own, your own model, at least start experimenting with it.

And then Cindy had a question around. Innovative ideas are not necessarily what I was thinking AI could help with. I was thinking about identifying potential partners. Can you speak to that? Yeah, maybe I'll take a first stab at where and then, and then feel free to add. So I think Cindy, I think it's a great question. I think, you know, when we work with, with corporations trying to implement a circular business model, identifying partners and specifically the people within those partners who are going to be participate in that business model is a massive challenge. Personally, I haven't seen the, an engine where we can just say hey, identify these partners for me. But I think maybe what we could think about is how do you generate in the same way that maybe we generate synthetic Personas to test a consumer idea against. You could potentially train the model to generate synthetic partners within a given ecosystem that you could then test your ideas against and identify maybe the pain points or the key opportunities that that those synthetic partners face and then use that as a way to filter narrow down your research as you pursue other part like real, real partners in your, in your research.

Yeah, I think there are a couple of questions around like now beyond innovation, innovative ideas, how do you see the AI to help on map the impact of a potential product could have on any of the ESG dimensions? So yeah, that's basically when we look into the development in the iteration stages. If you feed it with data and kind of, you put it towards a direction where you can start to evaluate the idea. So let's say you have in your company a certain, certain environmental goal that you have. Right. You can start using the model to kind of. Okay, so these are the goals that I have and they can be long again, these documents can be long and you can feed that information and it can use to kind of oh, according to your information, this is probably not going to be the best idea because of X, Y, Z. So maybe let's iterate on that direction. Right. So you can iterate on the ideas towards that goals. And then like related to another question was about the confidentiality. Yeah, that's. That's a critical element. So during this experimentation phase with GPT you need to be careful with it. If the environmental goals are kind of public information, then you probably can do that. But then you need to be careful on how you iterate on the idea inside the model that is public like GPT. But at least you can start experimenting with it. But many companies are starting to have their own kind of GPT space that is kind of siloed from the OpenAI or other. Yeah. Models.

Yep.

And on the innovative ideas is not necessarily what I was thinking. Yeah. On our. We had a. We had a summit last week around. Around. Yeah. Autonomous innovation. And there was an interesting study from. From a Harvard professor and that her initial assumption was that either AI would be better for analytical tasks rather than creative tasks. And after the study that she did in a consultancy she actually flipped that and that the impact of creative ideas were actually much higher than the analytical one, especially for these generative models. Because the generative models that are trained on language models and language models are not necessarily good on the very specific. Like if you ask a very complex mathematical question for it, it's going to get lost. But it's better on language. So everything that you have around numbers, it's also important to translate that into verbal elements so it can better iterate with you. So for instance. Yeah, the environmental goals like translate that into kind of regular language, like not mathematical language. So you can have, you can iterate it better with it.

Vinicius has a question of do you believe that AI will solve problems without human beings being aware of themselves and their impact on the world? Which is a very deep. I feel like a deep. An important, an important question. I think the how we define solving a problem maybe is a clear as a key definition or parameter for us to think about. I think we're seeing a trend towards more autonomous product development and leveraging consumer problems to inform a product development and then launching that. I think that will be in fact the. The case. We're actually just to do a small plug for. For next month. Etor shared a bit about this Stingray model which is kind of like our replacement to the double diamond approach from around design thinking. But as that Stingray part of that Stingray model implies is like human intervention as we go into testing. Right. So there's always I think going to be a level and perhaps should you know, this goes back to designing with, with intent and designing the parameters. But there should be a level of human intervention to ensure that these, these models don't get out of hand and we don't get into these to these shine situations with the swastika. Right. So I think like probably technically the answer is yes to your question in my, in my mind but potentially there's ways to, to, to, to prevent that or have more human intervention to ensure that there be their. The growth is done sustainably and ethically.

Yeah, we always use the expression copilot. Right. When we are talking about AI at least on this early stages would definitely be a copilot. And as we move on into the becoming more sophisticated around this then it's really about you know, adjusting this modeling this model. And of course as you get later on in the interation cycle then it becomes, yeah then becomes the human intervention can becomes more important but the, the AI allows you to do like in the breath and the amount of, of ideas that you can iterate on them on sustainability and desirability and all the different lenses like it really allows us to really go for a lot of them and then as you fine tune the humans start to get in more active.

Yeah. And I think this question. Erwan and Kirsten, I'm going to try to tackle your questions simultaneously. So Erwin asked if we've generated groundbreaking circular innovation with clients based on this approach or are we just at the beginning of the journey? And then Kirsten asked or mentioned that like challenges around sustainability tend to be extremely complex, messy systematic problems that and that like AI maybe we're conveying it as a, as a silver silver bullet. So the I think the short answer for both of those questions. Well or the answer for both of those questions in my view is that designing for circularity, designing for sustainability is incredibly complex because we're not just relying on launching a specific product. Right. That product needs to be part of a broader ecosystem that brings it up the value chain again. And so from, from that vantage point, AI is not a, A, a silver bull. We think that this is the worst that AI is going to be and helping us to come up with solutions that that can can solve those challenges and, but it's not a silver bullet currently. So Erwin, that question around like have we done this? Like specifically we've launched products that are maybe more attuned for circularity but this ecosystem approach that Kirsten is referencing, I Think requires more than just leveraging ChatGPT and other AI tools. It requires that human connection I systems thinking etc so we're on the journey towards doing that. But it's not a, it's not a. Currently a silver, a silver bullet for really achieving that full fledged level of circularity. Yeah, we're like one step into a do. I'd say it's a marathon. Let's see in the final minutes I'm trying to see if there's any other questions that we missed. And then I'll check the chat. Yeah, I think so. So one thing I would like to just reserve the last, last few minutes for is if you have ideas or areas of interest as we think about this idea of gen AI times and sustainability and circularity in that intersection. If you've found anything interesting today or there's specific topics that we didn't address or things that you want us to go deeper on, would be super curious and interested if you all could drop those ideas or topics in the chat and potentially we can have another session like this where we share our point of view but also have a chat dialogue around different opportunities, different areas for, for improving how we go about developing products for a circular economy. So drop any ideas you have there in the chat and if you have other questions that we missed.

There was one coming in the chat now how AI can enhance individual consciousness. Vinicius is coming in with a very philosophical questions here. No, yeah, this one is a very tricky one and I think taking it as, yeah, taking it as a co pilot is very important on that process and having someone to talk to, even if it's virtually in the first steps, I think it can help on this process.

Do we have any ideas to become a more design minded organization to facilitate innovation? Yeah, I mean I think fundamentally that's, I would say that's at the core of who we, who we are today. Right. Is intended to partner with clients to design products and services that fit the needs for tomorrow. And so that's our goal is to partner with the most ambitious organizations and develop what life needs next. So hopefully that's certainly the part of our ethos as a business. Okay everybody, I think that that covers it, please. I like the questions or the topics coming in so I really appreciate the feedback and the, and the ideas for, for how we could take this to the next level. As you, if you are curious like feel free to, to go to our website as well and lots of different topics around how to use AI for generating products and services. Some of them are specific to circularity. Some of them are not. But there's still, I think, elements of inspiration and tactical steps that can be learned and channeled for your own product development and innovation goals. And then, as always, like, feel free to reach out with any questions questions. If you're interested in having us come and speak about these topics or autonomous innovation more broadly to your leadership group, we have a team of individuals who do that quite often and happy to organize something like that with you as well.