AI and the Knowledge Economy with Rita McGrath and Sangeet Paul Choudary

If we keep on putting on that task centric lens and looking at, well, what does Kenny automate? We're going to keep, you know, staying stuck in the frames that we have today. Today's jobs, today's workflows, we'll end up eating them, making them cheaper, better faster. But we will miss out on the, on asking this question of why does this job exist in the first place? Why does this workflow exist? Why does the logic of the firm exist? Welcome to Thoughts Parks. Like, subscribe, comment, share and enjoy. Well, hello everybody. Read a McGrath here for another edition of the Thoughts Parks podcast where I invite really interesting authors and instructors and people who I think have interesting things to share with the world to talk about their work and their perspectives and what we can learn from them. And I'm absolutely delighted, my guest this week is Sengie Chowgary, who I met in person, in real life, at the Thinkers 50 where he received the Strategy Award. He's got all kinds of acclimations, all kinds of books. He is really considered an expert in platform strategies and has just published this very cool book called ReShuffle, which is quite generous and quite heavy. And as you can see, quite well annotated by myself. And the book is really about how do we think systemically about the changes that are happening as AI sort of filters its way through our economy. And I think has some very interesting and provocative things to think about. So welcome, welcome, Sengie, to the show. Thank you, Rita. I'm really looking forward to this conversation. Yeah, me too. So let's start with something maybe a little provocative. In the book, you use a historical illustration of the Maginot line. I think that's how you pronounce it, where France determined they were not going to be vulnerable to the things they felt vulnerable to in World War One. And they built this incredible fortified wall, basically, that proved to be ultimately not very effective. And your argument is that what a lot of companies are doing right now with respect to AI kind of resembles that. So maybe start there. Yeah, I think the story of the Maginot line is, it's a really interesting illustration to kind of talk about how we often end up creating a really elegant solution to a problem that itself has become irrelevant. And that's because we're using a left and stream that does not apply anymore. And so the story essentially is from the second World War, where at the end of the first World War, France decided they did not want all of that loss of life again that the first World War had brought. And so they created this series of fortifications along its borders. And the only part where they did not create it was where it borders into the Adiniv Forest, which was considered impossible. And when World War II happened, so these series of fortifications were supposed to be impenetrable. And they were impenetrable. So they were a really good solution. But the context had shifted because radio communication had been invented. And Germany had come up with this new form of warfare called the Blitzkrieg, where you could coordinate across multiple different forms of attack. And that allowed them to move very swiftly through the forest, through the swamp and attack France from the other side. So the key idea there is that very often we think not just about competition, but about all of our frameworks, you know, in terms of frames of thinking that no longer apply. And that's really my point over here, where in today's world, when we think about AI, we really often put on an automation lens. And that automation lens comes from the impact of technologies on our work for, you know, for ever since it's since the beginning of time, since the beginning of innovation and technology. But one thing we miss out on is the fact that over the last 25 years in particular, our world has gotten more connected. It has gotten more data rich and intelligent. And what that ends up doing is that today businesses are much more interdependent. They do not operate simply in markets they operate and connected ecosystems. And so we need to when we think about the impact of any technology today, we need to think about it not just at the level of its impact on automation of individual tasks that we perform, but also on its impact on the coordination between these different actors at a systemic level. So the key idea that I talk about in the book is that AI's impact is not at the level of automation of the task. It's also at the level of changing how the different parts of the economy connect. And that plays out not just at the level of, you know, different companies working with each other, but even at the level of work flows different parts of the organization working with each other. If we do not think about the impact of AI at the level of a system and do not think about its impact on coordination and whether it improves coordination or it solves previously uncoordinated, you know, it lets previously uncoordinated domains coordinate for the first time unless we think about those effects, we're really missing out on the real impact of AI. So if we want to think about what AI is going to do to us, if we keep on putting on that task centric lens and looking at, well, what does Kenny automate? We're going to keep, you know, staying stuck in the frames that we have today today's jobs, today's workflows, we end up speeding them, making them cheaper, better faster, but we will miss out on the, on asking this question of why does this job exist in the first place? All of that is up for grabs in today's world and that's where the real impact of AI plays out. Well, I really appreciated about your book and it's, it's a fascinating read, like just lots of really rich examples, but that you really do take a systems approach to the subjects that you're analyzing and very few books do that. You know, and I think, you know, there was this recent MIT study that basically said that something like 95% of all executives reported being disappointed in their implementation of AI and I'm like, of course, you know, what are they doing? They're sort of clinging AI onto the back end of something and hoping that's going to make a transformational difference. Well, no, you actually have to change the system. So, so perhaps before we dig into the next example, because I do want to talk about shipping containers, but for those of you just joining us, I'm here with Sanghi Chatteri talking about his very interesting book, ReShuffle, which is one of the more insightful takes I've seen on how AI is likely to have an impact on our world and certainly on the field of strategy. So what caused you to write this book? Where did it come from? Well, you know, it's sort of, so I think the trigger came from two ends. One was that I was spending a lot of time working with executives, helping them think about strategy and competitive advantage. And my work on this, you know, kind of stems from my earlier work on platforms and ecosystems and increasingly a lot of the questions that they were asking were about AI, but they were stuck in this model of well, how do we get higher productivity? How do we think about optimizing something? They were not thinking about what does this mean for the future of competition and our industry, what fundamentally new forms of firms will emerge because of AI. And so I wanted to answer those questions at the same time. The other trigger was from the many dinner conversations I was having with anybody and everybody who had an opinion on AI and this, a freak, you know, every conversation used to go back to this point, AI won't take your job, but someone using AI will. And so it was these two. Exactly. I was telling you, I was saying we were chatting, getting ready for this conversation about this hilarious cartoon that was sort of circulating on LinkedIn. And it shows a farmer talking to a horse and the farmer says, don't, don't, you know, a tractor is not going to replace you, but someone who, but a horse who knows how to drive a tractor. That is funny. Yeah. So it was these two fairly diverse triggers. The reason I landed on the systems approach was that if you look at it from a systems approach, you start seeing that the executive question of what does AI mean for my firm and the dinner side question of what does AI mean for my job? Both of them can can be answered very effectively if you stop looking at tasks and start looking at, you know, where does my job sit in the larger system? Where does my firm sit in the larger system? Yeah. And I think, so what my own PhD was from the social system sciences department at the Wharton School. And even back then, the big discussion was, well, you can't understand a system by decomposigated to its component parts and trying to understand each part. You have to really see how the whole thing operates together and very few business thinkers look at things that way. So let's go back in history. And you've talked about the shipping container as something that was fundamentally transported in. A lot of people just take this for granted, right? But if I go all the way back, you know, long, long time ago, if you look at the theory of, well, it used to be called internationalization and then it was called globalization. And there was a set of hard and fast rules about what kinds of products could be you know, usefully shipped around the world and what kinds couldn't. And the prevailing wisdom at the time, because they were operating within the constraints of the system, which was bulk packing and loading and, you know, dock hands and ports, and it would take days, you know, to get a ship in and unloaded. And I mean, it was just a really cool key system. And so the logic coming from that reality was that the only things that really made sense to export in large numbers were very high value items that you could get a huge prize premium for. La la la la la la. And the idea of shipping, you know, cheap furniture around the world would have been completely contrary to that logic at the time. Enter the shipping container and the standardization that it allowed. And all of a sudden, you've got these incredibly efficient global systems, which allow you to ship basically anything around the world for a very affordable cost. And that totally changed the dynamics of global trade, whose manufacturer and what, who's trading with whom. And it was, you know, I think AI has a lot of parallels to that in today's world. So perhaps talk about what that analogy might look like today. And I'm particularly thinking of when you talk about the mechanisms for coordinating tacit knowledge, especially. Yeah, absolutely. You know, the reason I took the shipping container example is people don't normally associate it with something like AI because it's not automation. It's not intelligent technology. And the other reason I wanted to take it was to show that the effects of automation are largely local. To your point, when the shipping container came in the first effects of automation, the first effects of the container were in poor automation when cranes came in and took the dock workers jobs. And today, when we think about AI's impact, we often think in those local terms. We're always thinking about which tasks are going to get automated. And to what you mentioned beyond that about the standardization, the real impact of the shipping container was they played out when trucks, trains and ships agreed to a common format of the container. They agreed to a common contract so that you could now ship from source to destination without needing to, you know, coordinate multiple times and manage multiple handoffs with systems that did not talk to each other. They were now fully coordinated. And the reason I feel that's relevant to where we are with AI today is because a lot of the knowledge economy relies on a tacit knowledge. So knowledge buried in our heads or expressions of that knowledge buried in unstructured formats in our emails and our phone conversations. But it also relies on a lot of knowledge that's just difficult to extract value from the cost of extracting value from a lot of this unstructured knowledge is so high that we just don't go about doing it. And so even if there is some value to that knowledge, we never end up uncovering that. Well, what AI does is it does do things. The first thing it does is that it makes sense of unstructured knowledge. So it's able to infer knowledge from unstructured data and it's able to make it usable, retrievable, most importantly applicable because you can run entire workflows around that knowledge. You're no longer just dealing with information which then has to be fed into a workflow where there are additional steps involved and they're again, you know, there's additional coordination involved. So you're on the one hand getting value out of unstructured data, you're inferring important value out of it. And then you're creating or you are making it applicable as you know, applicable knowledge and not just information. And so what that ends up doing is that progressively as AI trains on all of the unstructured data that's available, you know, with that individually in our enterprises on the internet at large, it's what I call unbundling some of our tacit knowledge, increasing a lot of our tacit knowledge and absorbing it into the tool. Now the reason that's important is that once you start seeing more of the tacit knowledge get absorbed in the tool, you and the fact that you can deal with unstructured information, you have the ability to start speaking a common language across different systems. So I'll give a very simple example to make this real. In the past, if you wanted to create a hotel rating system, you had to build a trip advisor. Somebody would have to go and infrastructure information and the form of a rating and trip advisor would compute that. Today, you could use AI to just work through all the discussions on Reddit, all the blogs, all the Twitter feeds and create a score for a hotel based on all of that information and determine for different types of use cases how well it's been rated. And that's just a really tactical example. There are many other ways you could think about it. But the fact that you could take all of that information which could not be used in the past, make it usable, make it immediately applicable, that is what helps drive the value of coordination today because there's a lot of information across systems which are not able to talk to each other unless they are structured, unless they have, you know, way structured interfaces to communicate through AI, solve a lot of those problems by working with unstructured information and then feeding them into workflows which again span multiple systems. I think that's absolutely fascinating. But the other thing that I think is, that you touched on, is so what does this mean for jobs and roles? And what I think we don't appreciate enough is that, you know, a lot of jobs exist because there's some kind of friction or inefficiency or, you know, scarcity in the system. And once that gets eliminated, that job is no longer really necessary. And the three that you mentioned were, so you got typists, you know, I remember there used to be a group, I think I'm young enough that it was no longer called typing, it was still called word processing, but there was a word processing yet. And all documents came and went from that unit and they sort of did it. And then you got telegraph operators, right? And those used to be very highly trained, you know, people that had to do that. And then loggers, right? Which once, once, you know, the automated logging equipment came in in the form of chainsaws, you know, the deep skills of the loggers were no longer sort of necessary. And so it gets to one of the fundamental characteristics of disruption, which is it takes something that was really complicated and makes it simple so that more people can engage in it. And it takes something that used to be really expensive and really inaccessible cheaper or, you know, affordable and accessible. And I think what a lot of people don't understand is that yes, that's going to have an immensely destructive effect on jobs that were only there because there was some kind of scarcity, but it's also going to create demand in other places. Now the trouble is to me anyway, at a societal level, there's a disequilibrium in terms of time. So destruction tends to happen first. And then the reconstruction happens afterwards. But I think a lot of people are just looking at it too narrowly, they're like, oh, is this going to take away my job? Well, sometimes the answer is going to be yes, you know, if you're 55 year old translator, you know, operating in, you know, Belgium or something, you know, what are you going to do? Because I can do so much of what is necessary. I mean, you might still need to put the final touches of expertise on at the end, but it's not going to require the whole job, right? So I think people don't appreciate that enough. You make a big point in the book that a coordination is one of the, I think one of the magical potential properties of AI and you specifically play out, how do you coordinate test technology, observe the world, create a working model, reasons through possible choices, acts on decisions, and then the critical thing is different about AI than a lot of other technologies. It actually learns from its past, which I think is just fascinating. So I just thought those were really, really, to me, very, very interesting points. And so let's shift a little bit to what does this all mean for strategy? Because, you know, I go back at one point I was thinking of doing a strategy textbook, so I was reviewing a lot of existing strategy textbooks and I'm just looking at them and it's like going backward in time. Yeah, you know, I mean, that's, that is my central fascination, not just with this book, but with a lot of my work, how is strategy changing? And I believe that's where we have a lot of common ground as well. So, you know, my fundamental view and I would like to then bring AI back into that frame is that a lot of strategy that we, you know, a lot of the work on strategy that was done or the frameworks and strategy that were developed through the 70s to the early 2000s were built for a world where the structure of industries was largely stable. And if I were to use it more broadly, the structure of any system was largely stable. In that world, we largely dealt with operational uncertainty. So fluctuations in demand, supply shocks, these were operational uncertainty. They did not change the structure. Today we deal with what I believe is structural uncertainty, which means that the structure of industries are for grabs. And there are many reasons for that, but just a very simple example, you know, the boundaries between industries are no longer static and fixed. And especially with AI, you could train a model with knowledge in one industry and make that model portable to another industry. So the boundaries between those industries has completely collapsed. And this was happening with data, but we were using it only for structured sensible data. Now we can do the same thing for all kinds of tacit knowledge. And that's really what's exciting about this because it applies to nearly every aspect of the knowledge economy. You can make knowledge portable and move it away from the place where the investments and the learning for that knowledge was originally created and make it usable and applicable in other domains. So the fact that boundaries of industry that changing, the nature of tasks that create value, which task hold value versus which do not, all of these things are changing because going back to your point, if something becomes accessible and widely available, it becomes less scarce. And so economically, it's value falls. So value no longer sits at that part of the value chain or value stack, whatever you call it, it moves to another activity. So all of these things open up a world of structural uncertainty. And the reason this shift from operational uncertainty to structural uncertainty is important is because in operational uncertainty with a stable playing field, we delight a lot on what is called positional strategy. You, you, you know, you're the light on occupying a specific position. You developed competences to defend that position. And then, you know, even if you had to evolve your capabilities, you did so within the assumptions of a stable playing field. Today, the dominant way to win is not to play a game better by, you know, not to play a static game better based on your position. The dominant way to win is to change the rules of the playing field for everybody else. And I believe that's the main shift in strategy that's happening that your, you know, the, the whole idea of a stable playing field with the position that can be defended is going away. And there's a divergence between certain firms who are able to restructure the shape of the playing field and the rules of the game within it and other firms who are forced to play that game within that playing field. And that's really the difference that's coming up. We've seen that, you know, with the rise of algorithms and big tech. We've seen that in many other cases, you know, the Alliance, G-O-N-N-D-A-C-M-O-D-I-Zing, the traditional telco model and, you know, restructuring the telco industry. So you've seen it across the board, but with AI, it's going to become faster than ever. So in my book, you know, just to wrap this up, in my book, I talk about this fact that I talk about this idea that we used to think of strategy as, you know, going back to Lafayette Martin, the two questions, where to play and how to win in a world of operational uncertainty, but stable structure, where to play with markets, industries, customer segments, geographies, and how to win was positional. But in a world of structural uncertainty, where to play is about how do you structure the playing field for everybody else? What are the constraints in today's playing field which you can solve? And because you solve it, you create gravity towards yourself. As the web is exploding, whoever solves search creates the gravity towards that. You know, as mobility is coming up, whoever creates a platform for apps to be organized creates gravity towards that. So there are players who set the rules of the playing field and there's everybody else who plays within those rules. And that's the divergence that we are seeing today. Yeah. And I thought the focus in your book on constraints and this fascinating question of, you know, there are things people do that have incredible value, but which nobody's prepared to pay for. Right. And so you mentioned these guides right across the desert, incredibly valuable, but they don't get to capture very much of that value. Care jobs, you know, incredibly valuable, but we don't pay care, carers, you know, very much. And this idea that the constraints, and you spend a lot of time talking about scarcity, risk and coordination. And then the idea that if you are adding value, people have to understand what that is, or nobody's going to be going to pay for it. And I think this is a really subtle, but incredibly interesting concept, which is value is not commensurate with monetization potential. And that just struck me as, wow, you know, we don't teach that enough. I don't think so. Yeah, absolutely. And just to make that point land home for the reader, I made the distinction between intrinsic value, which is the inherent value of doing something, economic value, which is how scarce is your ability to do that? And hence, you know, how it's priced in the market. And contextual value, which means that every activity or capability has value in a certain system. If the logic of the system changes, how that capability is valued will change as well. And so with AI, especially both these things play a role because AI changes the economic value of certain capabilities by absorbing them into the tool and making it widely available for everyone to use. So translation is a great example of that economically valuable in the past, not economically valuable as much anymore, because you have access to AI translation. And yet, certain forms of translation still hold contextual value. So if you are in a high stakes negotiation, you want to ensure that you get the translation absolutely right and based on the tone and nuance in the room. And so that's where a human translator still has a strong position. So making that distinction between, while the task of translation still has value in the sense that it still creates value for us, its economic value has changed. And only in certain cases, it's able to monetize because it retains contextual value. So that's the distinction I wanted to clarify because when we think about AI very often, there's a cam that reacts, the human touch will always have value. And they're talking about an intrinsic value. Yes, it will have value. But you also need to think about economic value and contextual value alongside that. Yeah, absolutely. The other thing that I think is interesting is you have a number of examples of firms that have just kind of rethought the taken for granted assumptions of how you compete in this market. Right. And one of the more interesting examples, I think, is that of TikTok. So social media grew up with this idea that you sign onto social media or dating sites or whatever because that's where other people are. And I think you refer to that as the cold start problem, which is, well, you know, chicken and egg, right? I was like, nobody's on the social media site. Well, I should I sign out of it. And then in conjunction with that is this idea that once you got all those users, right, that, oh, you're going to have these powerful network effects and that's going to make you, you know, a domatable competitor and nobody else who is going to win. And so along comes TikTok and they said, well, actually, we'll build an algorithm that can simply go off your interests. And so whether you have one connection or a million, it doesn't really matter to us. We're going to feed you, we're going to show you what our algorithm determines you might be interested in rather than what your friends are doing. And it, you know, that sounds so simple and yet it is completely transformed to social media landscape. Right. So I think companies that are able to kind of really rethink what are the taken for granted assumptions we're operating on? And why, you know, well, I really ask that base question. And I think that's just such an interesting approach to thinking about, well, what are the constraints in our strategy that we can actually work around? Yeah, yeah, absolutely. I think, you know, the TikTok example is so easy to the late two. A lot of people see TikTok, but they just assume it's, it became successful because of its features or, you know, a new way to do habit design. But really the distinction over there is that both Instagram and TikTok had access to the same technology. Instagram used AI to improve recommendations within the logic of the social graph, which is you get to see stuff on your feet based on who you're following. And now those recommendations are better tuned, but it's still within that logic. And TikTok essentially eliminated the need for a social graph. And hence was able to bootstrap from scratch while others had dominant network effects because Wall Street at the time believed that was impossible to unseat Instagram. And so this idea of, you know, taking learning and creating an intelligence first network rather than a social graph first network is an illustration of, you know, how the roots of competition can change with AI. And I'll give her with you, you know, another interesting, the later example because we often think of TikTok and we say, well, that's fine, it's digital. But the same thing has happened in fashion. And I take the example of Shane, the Chinese Fast Fashion Company. Again, you know, Shane is delighted and is known for everything except the business model innovation that it's done because of AI. Old fashion or, you know, traditional fashion used to work on this model of a seasonal cycle where you would fly a merchandiser. to Milan or Tokyo, they would figure out what styles are emerging. They would come back, create the designs, run a big batch, and then hit the market with the full seasonal batch. And what Shane does is it completely changes the rate at which the cycle works, but also what's really driving the cycle. So Shane constantly captures data from TikTok, Instagram, and other social media. It determines micro trends, and for every micro trend, it has this, it has a network of factories that can build out a very small batch. It then tests that batch in the market. If it works, it doubles sound on it. If it doesn't work, it moves to the next trend. So it's created a learning machine to capture most of the tacit knowledge and fashion. So a typical designer on Shane is really just doing one or two tweaks on an idea that has been largely algorithmically designed and curated, whereas in traditional fashion, all of design was supposed to be something that machines could not do. So it's just interesting how by changing the nature of the cycle and the speed at which it can learn from the market, it created a fundamentally different approach to fashion. And that's something that we're increasingly seeing happening across industries. And it's interesting to me as well that I teach it to business school and as do you, you know, at Eard Dartmouth, right, in a few other places. And there was a time there where you could not make it down the hallway without tripping over case studies of Zara, and these fast fashion companies that had completely unseeded the traditional fashion houses. And yet, you know, here's Shane. Yeah, Zara, I know what you think is really, really interesting. You also talk a bit about Sephora and what you mean, what a bunch of those companies are doing is sort of simplifying decision making for the consumer. And I think that's really interesting too because when I work with companies, it's very typical that, and I understand it, they start with what they do. And that's the way they're throwing you world. And they really don't get out into the customer's mind and go back and say, "Well, I call it a consumption chain. You know, where does chain break down? Like, where do customers get stuck? Where do they not able to move forward?" And one of the things that AI gives us is the ability to much more richly understand those moments, you know, when customers are confused or when they're searching for an alternative and they know they have a problem, they just don't know that even that there is a solution. And companies can create a really great strategy around, you know, making that whole chain just work better. So maybe talk about Sephora a little bit, I think that would be fun. Yeah, absolutely. I think, you know, one of the key ideas over there is that when we think about the knowledge economy, we always think about the knowledge work we perform as producers, but every consumer going through a customer journey is also performing knowledge work. And AI has the ability to change how and where that knowledge work is performed as well. Whether it's a consumer or it's an influencer to the consumer, they're all performing that knowledge work. And with AI, the way that gets performed, how often it gets performed and for whom and how widely it gets performed. And hence changes, you know, the definition of your target market, all of that changes. So I take the Sephora example because it's a fairly straightforward illustration of the fact that even though we think of Sephora as being in the business of selling cosmetics, it's really in the business of managing or the do you think uncertainty about how you should manage your beauty regimen, how you should take care of your skin, your hair. And in order to do that, it's real, you know, it's real play today is as an assistant guiding a consumer to those choices. So if you walk into a store, they have, you know, a scanner based on which they determine the pigmentation of your skin. And on the basis of that, they assign a color IQ, a data profile to you. And that starts training Sephora's algorithms about what to, what form of regimen would work best for you and which cosmetics would fit that best. And so the key idea with all of this is that consumers are, you know, dealing with uncertainty in their decisions in terms of, you know, what to buy, what's the right thing that's, what's the right product for their need. And even at the time of usage, how best to use the product. And AI is a way to solve that problem at scale because today, the problem to a large extent is solved by human intervention in various parts. So whether it's a buyer in a consumer market, talking to a store assistant or a buyer in an enterprise market, working with a reseller, all of that knowledge can be used to train them. And AI assistant, which can be provided to a customer at the right point in the journey. And when I say an AI assistant is just insertion of the right amount of intelligence at the right point in the customer journey to solve that customer's problem. And we've already seen, you know, 15 years, a 10 to 15 years of whoever owns the customer interface has a strong position in the ecosystem. And with AI, I think it's going to become even more accelerated because you will have many different players working with many different forms of intelligence that they can insert into the customer journey. So I think that's going to become increasingly more of a competitive battleground in the years ahead. Yeah. Yeah, I would agree. So sort of toward the last part of the book, you talk about the emergence of, I'll call it, two generic sort of roles you can play in a sector, I'll call it. One is being a solution provider, so offering something directly to the customer. But then there's also fascinating emergence of the tools people. So, and I thought a very interesting discussion about, it's obvious that one is better than another, but each of them have like their advantages and disadvantages. And with the tools people, I'm very reminded of Cory Doctoro's very famous phrase and shittification. And he talks about, in the beginning, you create your tool to be absolutely delightful and everybody gets on it. And this is great. And once you've got a certain amount of power, then you kind of, you have the customer sort of locked in at some level and then you can capture more and more of the value of yourself and possibly degrade the customer experience. And to some extent, I think that's a little bit of what's going on with Amazon right now. I mean, if I go to Amazon and I take the, you know, "Quees in Art Toaster," I get like five pages of advertisements for other people's toasters before I finally get what I want. And so I think that to me is a great example of how, and I understand why that might be economically attracted to them. But at some level, somebody's going to come along and say, "Well, you want to quees in Art Toaster?" We can just get you that. And then the advantage shifts, right? So it may be talk about the difference between tools, the danger of becoming overly reliant on somebody else's tool. And then what are some of the advantages that solution providers have as opposed to the tools people? Yeah, I think this is going to be a central tension. And it already is showing up as a central tension in the age of AI, where you have at least, you know, at this point, anybody who's providing the models and the attendant capabilities, they are providing the tools on top of which professional services companies and even other software companies are providing ends solutions to their clients. But as we've seen over the last several months, you can have the tool providers either move upwards and capture important parts of the solution provisioning. So a classic example is chat GPD started as this, you know, completely, you know, simple chat application, which you can ask any question. And you had all the individual players, whether it's an e-commerce company or a consulting company, all plays in different industries using the models. But increasingly, chat GPD is moving into those industries. You know, there's a talk about chat GPD becoming the e-commerce interface where you start your customer journey going back to the previous question as well. Instead of having to click through and work through different options on a website and you know, manage that whole coordination of am I buying the right thing? Chat GPD guides you through that whole thing and you end up buying the product right within its experience. Two, you know, talks of chat GPD getting into consulting as well. The point is that what started as a tool and as a horizontal tool that was not specific to solving a specific problem for a specific customer and hence the solution providers who were using it, who were using it as a back end capability to serve their customers, they started out by not feeling that threat, but I increasingly think that threat. And we've seen this before, you know, I take the example of the relationship between Uber and Google where Uber uses Google as a tool provider because Google provides the mapping capabilities. But as Google has learned from everybody using the maps, it's now moved, you know, through a bit and way more it's moved into providing the end solution as a light-hating service as well. So we are going to see that increasingly with today's model providers who are today's tool provider as well, moving further into the solution side. The solution providers who are building on top of that trying to reassert value because the thing is that if a tool climbs up into the solution, the margin that the solution provider has gets increasingly compressed. You could say that as a consulting company, your productivity is increasing dramatically with AI and so you need fewer hours and hence your margin is higher. But then either some of the work will move into the client because the client can use the tool directly or some of the margin will be captured by the tool because your productivity is so dependent on them. So the margin compression is bound to happen as this place out. The key difference between a solution and a tool though is that a solution solves all the constraints involved in getting the customer's outcome. So a solution provider will increasingly have to move closer and closer to the outcome. You can't just provide capabilities and charge for that. In many ways today when we in consulting or doing that, time and materials billing, whoever is doing that is providing a capability and charging for that. That model starts going away and you have to move closer to owning the outcome to delivering the end-to-end solution to truly qualify as a solution provider. You also have to increasingly own the risk and manage the risk on behalf of the customer to significantly distinguish yourself from a tool that's getting more powerful underneath you. So there are all these tensions that are increasingly going to show up as AI gets used more broadly in the economy. Yeah, it's really interesting. And I like the way you categorize it, that rather than thinking about industry. So what do you think is going to happen with consulting? It's an interesting question. It's still evolving. I do think that consulting is not a one-size-fits-all solution and that the way consultants compete or differentiate themselves is not the same across the board. So at a very high level, there's some consulting that is relationship-based and you're not really hiding brains. You're just hiding continuity. And those consultants, I believe, if what they're really selling is continuity, they might have to tweak how they monetize it, but because the primary value they were providing was continuity and now they have better brains to support alongside that, they might actually benefit from it. On the other side, if the primary value of providing was we higher the best brains, we train them so that you don't have to and you can hire them for a fee at any point, they might feel increasing amounts of margin compression. And so depending on what you're actually selling, I think it's going to become very important to unbundle to use the terms that I use in the book to really unbundle the propositions that you're selling to the client. Ask yourself, where do you really have a right to play? Is that getting attacked by AI or is that getting enhanced by AI? Based on that, if it's getting attacked, you might have to migrate to something else and re-bundle other sources of value that are still not or that are relatively immune to being attacked by AI. And so I think that's the key. That's why I make the distinction between selling continuity versus selling intelligence versus in other cases selling certainty because you are managing risk and you are charging on the, you are doing a revenue share. So depending on what you're selling, how this affects you might vary dramatically even though you might, even though all of them might be consultants. Right, right. Yeah, I mean, I think it's it's it's going to be super interesting how that all plays out. I mean, I think there's there certainly the relationship based consultancies, the, you know, McKinsey said this was okay to do so, you know, I absolved myself of any responsibility. I mean, there's still going to be a need for that sort of thing. But I think it's very different. You also talk about different business models, which I think is interesting in terms of the re-bundling, the building blocks. So the idea of things being rentable, things being recombinable and things being scalable. And again, I hadn't really thought about it that way until the new book, which is, you know, Mr. Beast and the burgers, although that was fascinating. So he saw, you know, decides to go into burgers. He's not a restaurant guy. You know, rents all these pop-up restaurants and millions of people want to want to buy Mr. Beast burger because he's got that to be footprint of users. And yet the PC doesn't control as the quality of the end product. And so I thought that was just interesting that he'd sell pieces of the business model problem, but not the whole thing. Yeah, I think that's, you know, that's already, we've already seen that in parts of the economy, where capabilities are available modularily. I also take the example of, you know, Uber users, Twilio to manage the messages that come to us. And either Uber or Lyft used Stripe to manage the payment. So we're living in this, you know, building blocks economy, if you will, or interdependent activity, the economy, where different parts of what you provide are can be sourced as a building block from third-party providers. What you want to do though is you want to know what you're really selling. Goes back to, you know, are you selling continuity versus intelligence versus certainty? But be very clear about what you're really selling and ensure that that's not coming from a commoditized building block. That's where you have a unique site to play. And that's the idea of the bundling. So with AI, this becomes even more interesting because at this point, it's only certain forms of process capabilities, you know, the ability to process payments, the ability to process how a message gets to your phone, cloud communications. Only those things have been modularized. But with AI, any form of intelligence is now modularized, which means that when you create a solution, some parts of the solution, certain parts of the knowledge components of the solution are coming from openly available capabilities. And so unless you have a unique way to bring it together and create a unique product on top of that, you might end up just playing with the commoditized pieces. So it's it's both a period of great opportunity because the kinds of knowledge work that you could access only by hiding large teams in the past can now be available to solo-pronours. Literally anybody with a $20 subscription. But at the same time, what will differentiate you is what can you uniquely do with all these capabilities available at your disposal? What's your unique right to win just like Mr. Beast's Heather's own? What is your unique right to win? And can you control what you're really selling? So even for people who are thinking about jobs, you know, or thinking about the impact of AI, will AI take my job or will it make me better at it? Is really the long question the right question is what can I uniquely do given that there's this whole surge of capabilities that independent of me are improving on a daily basis just on account of the funding that's coming in and will keep improving as we move forward. What can I uniquely do with all of that available to me? Yeah and something I'm working on as part of my next book if if I forget this organized is you know once the unit of value creation becomes teams and even individuals which AI can facilitate. You know there are fewer and fewer contexts in which you actually need large, megalithic corporations that what you may actually be looking at is you know knitted together ecosystems each of each part of which we're doing a specific thing that is valuable but even in strategy the concept of a firm the concept of competitive advantage I almost think we need a new word. I'm not even sure what that is. So you mentioned to me that you're working on a new project as if this wasn't enough I mean this book's pretty pretty chunky. So for those of you that are joining us late I'm here with Sanghi Chattery and his recent book is called ReShuffle and it's a really great systems level look at what AI might mean for strategy for our economy and what it means for you and your jobs and the fundamental point of the book is that looking at AI just in terms of tasks that it can do is the wrong way to look at it. We really need to take a systems perspective but you said you're working on something new I'd love to hear about it. Yeah it builds up you know the previous answer or the previous discussion we had earlier in the earlier today on the difference between operational and structural uncertainties. So my next book is called Unfair Advantage and the key thesis over there is that traditionally advantage was about playing the existing game better with better capabilities and creating a more to the idea of self and protecting that but increasingly we're moving into a world where the structure of the playing field is up for grabs and so the most the best way to win is to actually shape the playing field for everybody else and you know some of the key ideas that it talks about is that just being good at sensing change being good at responding to it if all of it is developed at an operational level that's not so. sufficient. If the rules of the game, the boundary of your industry and the nature of advantages constantly changing, you have to think about it at the level of how do you fundamentally change the game whenever you have the opportunity to do that. And I take a whole range of examples from how Tesla worked through shifting the game in the EV industry starting from solving the constraint in batteries to solving the constraints and charging and then eventually opening that out and commodifying that. And many other examples of this sort which essentially show that the game that companies are playing is no longer static because the nature of the playing field itself is not and so the most valuable advantage today is to shape the game that you want everybody else to play. You become the rule-setter and they play within that game. So it's a bit of a potentially controversial idea still working on fleshing that out, that's the idea of the next book. Super interesting. Super interesting. Oh, I think you're onto something. I think this whole idea of the dynamics changing is really key. You know, something I've been playing around with is this idea that we're really in a transition, you know, kind of a one-synogeneration too, even transition from this world, trolleying mass production, huge hierarchies and big factories and all that. Two-world which is, you know, really digitally informed and in which services are replacing products, in which dematerialization is a real thing, right? So once, you know, once you've got music that is no longer a change to a physical artifact of some kind, you know, there's so much more you can do with it. But it also completely changes the economic. I mean, for the record labels, it was awesome to have people forced to buy 18 songs to get the one that they really wanted to on a piece of, you know, CD or something. And their profitability, their margins have really become compressed. It's become a different game for the artists. You know, the scarcity has now gone to the live performances. So it's not universally wonderful for everybody. But on the other hand, it's vastly increased music consumption. It's vastly increased the number of people who could become musicians and who could consume music. And so it's interesting to me how the dynamics play out for the different parties to the system. But I love the idea of reshaping the playing field. That's interesting. Because it's also not always obvious when that's happening. So you know, go ahead. Yeah, no, I think, you know, since you mentioned the music example, I think that's one of the most interesting examples of how to think about moving value and creating fundamentally new playing fields. Because if there's one person who, you know, there's one company that progressively benefited the most from the shift in the music industry from CDs to MP3, it is Apple. Because when Apple, you know, Steve Jobs wanted to launch the iPhone or wanted to launch a phone back in 2000, but realized that 3G connectivity and, you know, other complimentary capabilities were not available yet. And so he sends this opportunity in music to organize all of this unbundled music and gave most of the 99 cents back to the record labels. But in that, you know, in doing that, he commoditized music and moved the value into the iPod. So the iPod was, you know, you would pay much higher than you would ever pay for a walkman or for any other consumer electronics device. So it was not a consumer electronics device as much as it was a consumer electronics device integrated with an ecosystem of easily accessible music. And when he eventually launched the iPhone, all the value that he had accumulated in the iPod because of iTunes, he was able to move that into the iPhone on day one because he commodified the iPod as an app on the phone. And so iTunes could now, all the attendant value in the iTunes that the music industry had provided him could be moved to the iPhone on day one. And I think that's such an interesting way to, you know, the shape playing fields, taking value from the music industry, moving it into the consumer electronics industry and then moving it from there into the computing industry with which he was able to then reshape the playing field in cameras, in GPS, in many other spaces, all of which, you know, got collapsed into the computing space. So I think that's such a, you know, interesting way to completely change how competition works the way he did that. Yeah. The other really interesting thing to me about that story was in a lot of companies. You know, the iPod was a fantastically profitable product. And in a lot of companies with this operating with the conventional strategy mindset, you would never have done anything to capitalize that franchise. You know, they would call it, and depending on the incentives in your system, there'd be a whole array of managers who dead set to defend it. And yet they pretty smoothly made that transition. And well, the people, a lot of people have forgotten is that before there was the iPhone, there was a thing called the iPod touch. Do you remember that? Yeah. Yeah. The original iPod had the click wheel, right? And then they moved to other forms of user interface. And then they were experimenting with the touch screen, but they didn't do it with a phone yet. They tried it out with your music player. And as I recall, if you connected it to your computer, you could actually use apps on it. I owned a couple of them. At one point. And I thought they were correct. And then I really had to do is add the telecom piece to it. And now you have what's going to work in prototype. So I thought that that whole progression was very interesting in terms of, you know, not being afraid to cannibalize what had been successful in the past. And then built this. Yeah. Exactly. I think, you know, if I may just add one last line over there, that's a classic example of playing with the new framework in mind, because the traditional framework of cannibalization was used to apply to product level. So if you were cannibalizing a music device, you're bringing a better music device with better features. And when you think of, you know, a connected ecosystem from music to consumer electronics to computing, you're not simply just cannibalizing and trying to replace it wholesale. You're trying to also migrate the value that you accumulated there by commoditizing that location, but then moving it into the iPhone. So, you know, even if people see what's happening, if they say another music company would have done something similar and had applied just a product cannibalization mindset rather than the commoditization and value migration mindset, they would have missed what he had done. So that's again an example of, you know, thinking about what was being done with an older lens versus what will actually achieve in the newer lens. I love that. I think, well, I'm really looking forward to your new book. And I think, you know, where we are with strategy really has to evolve much more about, you know, really much more about the systems approach than, than I think traditional strategy looked at. And so the book is great. So I've been talking with Sanji Chaudhary, looking at his new book, ReShuffle. He's already got another one on the way. So you better read this one quickly before that one next one comes out. And just want to pleasure to share some ideas with you. Congratulations on the strategy award. That must have been a nice moment. And thank you so much for spending some time with us on this podcast. Thank you, Rita. It's been such a pleasure. I've been a student of your work and it's been an honor to be on the podcast. Thank you.