How Will AI Reshuffle Power in the Global Knowledge Economy?

HOST (Paolo Sironi): Welcome to Banker’s Bookshelf, where we explore big ideas shaping the future of banking and fintech. Today I’m joined by Sanjit Paul Chowdhury, author of the groundbreaking book Reshuffle: Who Wins When AI Restacks the Knowledge Economy?

Why Are Predictions About AI Often Misleading?

Paolo: Sanjit, many people swing between extreme optimism, AI solving all problems and ushering in universal basic income, and complete pessimism, AI destroying jobs. Why are both of these views incomplete?

Sanjit: These narratives are too narrow. We tend to analyze AI at the level of tasks, what a specific algorithm can automate, rather than at the level of systems. Every major technology reshapes the entire economic structure, not just individual jobs or processes

• We over-focus on AI’s “intelligence” instead of its ability to reorganize industries.
  
• Technology rarely just replaces tasks, it unbundles and re-bundles whole systems.
  
• Winners are those who understand and exploit these systemic shifts.
  

QUOTE: “AI will not simply replace people who don’t use it. It will reshape the entire system of work and value creation.” – Sanjit Paul Chowdhury

What Can History Teach Us About AI’s Systemic Impact?

Sanjit: To understand AI, look at unintelligent technologies like the shipping container and the barcode:

• Shipping Container Revolution: Initially seen as port automation, the true value emerged when ships, trucks, and trains adopted a single standard. This coordination collapsed transport costs and enabled globalization
  
• Barcode Transformation: Retailers viewed it as faster checkout. Walmart used it to centralize inventory data, flipping power dynamics with suppliers and creating new economies of scale.
  

These examples show that technology’s greatest impact comes from coordination and reconfiguration, not just efficiency gains.

QUOTE: “The barcode didn’t just speed checkout, it redistributed power across the retail ecosystem.” – Sanjit Paul Chowdhury

How Does AI Enable “Coordination Without Consensus”?

Traditional breakthroughs required consensus, like shipping lines agreeing on container sizes. AI changes the game by enabling coordination without consensus

• AI can extract and structure fragmented, unstructured information.
  
• This allows multiple stakeholders to act in concert without formal agreements.
  
• Examples range from travel planning with AI assistants to orchestrating decisions in complex industries like banking, healthcare, or construction.
  

QUOTE: “AI provides a mechanism to coordinate markets without consensus, unlocking power where no dominant player exists.” – Sanjit Paul Chowdhury

What Does This Mean for Banking and Fintech?

Paolo draws parallels to banking:

• In the 1990s, banks raced to build proprietary risk algorithms.
  
• True advantage came not from the best model but from integrating risk management across silos
  
• Similarly, AI’s power lies in orchestrating knowledge across the enterprise, not just deploying better algorithms.
  

This insight applies to core banking modernization as well. AI can reveal hidden business logic embedded in legacy code, but real advantage comes from keeping knowledge explicit and coordinating decision-making across the organization.

How Should Individuals Prepare for an AI-Reshaped Economy?

Sanjit: Jobs are built around constraints, scarcity of skill, risk management, or coordination. AI removes some constraints (like task scarcity) but creates new ones.

• Don’t just learn new tasks; identify where constraints are shifting.
  
• Focus on roles that manage risk and enable coordination across complex systems.
  
• Skills like time management, teamwork, and creativity will grow in value as systems become more complex
  

QUOTE: “Look for the constraint and where it’s moving. That’s where the next high-value jobs will emerge.” – Sanjit Paul Chowdhury

Key Themes and Insights

• System-Level Change: AI’s greatest impact will come from reorganizing industries and redistributing power.
  
• Power of Data Visibility: Like Walmart with barcodes, those who control system-wide data will dominate.
  
• New Skills for a New Era: Coordination, time management, and creativity will matter more than technical tasks.
  
• Banking Opportunity: Banks that leverage AI to orchestrate knowledge, not just automate tasks, will gain lasting competitive advantage.

Raw Transcript:

Very often when we talk about AI, we sort of go into two extremes. We think about AI changing everything and saving our world and solving all our problems and creating incredible prosperity and everybody will then go into universal basic income. And on the other hand, we talk about, you know, everybody losing their jobs and AI taking away jobs.

And we have these two, you know, extreme views, techno-optimism and complete doom and gloom. And neither of these views is actually completely correct. This week, we spotlight the newest show on the Provoked.fm network, Banker’s Bookshelf, hosted by Paolo Serroni.

In this episode, Paolo interviews Sanjit Paul Chowdhury, author of the new book, Reshuffle, Who Wins When AI Restacks the Knowledge Economy? Paolo and Sanjit discuss how AI is set to transform the global economy and company operations by redistributing power among key players. Sanjit shares research that explores four central tensions and draws on compelling examples, historical parallels, and innovative perspectives. They also uncover the opportunities and challenges that AI presents, as well as who stands to succeed or struggle in this reshuffled landscape.

Buongiorno a tutti and welcome to the Banker’s Bookshelf. As you all know, the Banker’s Bookshelf is our monthly opportunity to learn from the authors of the most interesting books and from the researchers who challenge our thinking about the banking present and its fintech future. And today, we welcome back a returning guest.

I’m sure you’re all uploading enthusiastically because Sanjit Chowdhury is truly one of a kind. Sanjit was our first guest on the Banker’s Bookshelf, joining us two years ago to discuss his best-selling book, Platform Revolution, and his latest research on sandwich economics. Now, these two thought leadership works have established him as a global authority on business strategy and digital economies.

But today, he’s back to share key insights from his brand new bestseller, Reshuffle, that I have here with me. Sanjit, welcome to the Banker’s Bookshelf. Thank you so much, Paolo.

Really looking forward to this. Sanjit, I read every time all of your writings, and this one is really a masterpiece, I say. It’s truly revealing, and I invite everyone after this show to get the book and to start it as I did.

And what struck me is that you sort of start saying that we all got it wrong, which is a provocative statement. And actually, I agree. In terms of thinking about the potential impact of AI, essentially, most of the researchers out there focus frantically around the definition of efficiency.

So they’re very task-focused, like macro-managing the application of AI, which also sparkles a debate about the periods of human replacement. So you were doing this, you won’t do this anymore, there’s no jobs in the future. And then the narrative started being complemented by a conversation about how to augment the way we work to stay competitive.

In a sense, it’s like saying AI will not replace jobs, will replace people that don’t work with AI. And you’re saying, no, that’s not correct, it’s not the right perspective. It is a restricted way to imagine the future of AI, and therefore that can lead to false or wrong investments in the application of AI.

So now, explain to us what is this core idea that Reshuffle promotes a new framework to look at the impact of AI in firms and across ecosystems? Yeah, thank you, Paolo. The core idea of Reshuffle, actually, if you really think about it, is not entirely new, and yet it’s very hidden. We’re not, you know, even though we see it with previous technological shifts, we sort of remain blind to it.

And the key idea is this, that very often when a new technology comes in, it’s especially true when we think about AI and how we are reading its effects. Very often, we think about it in terms of its impact on processes at the level of tasks. So a task that was previously being done by a human now gets done faster, cheaper, better.

The cost of skilling somebody and training somebody to perform the task can now be replaced because you now have AI that can do the task. And so very often, we think about the impact of technology in terms of its impact on, you know, something as basic as tasks. And the key idea that, you know, I’m bringing forth is that every technological shift has really played out not in terms of how it’s impacted specific tasks or specific points in a system.

It’s played out in how it’s transformed the entire system, how it’s sort of taken the old system away, broken it apart, and then reassembled the new system, what I call unbundling and re-bundling of the system. And so a few simple ways to think about it, because one of the things that I start the book with is this framing of why unintelligent AI matters. And the reason that’s important is because we are too focused on the intelligence part of artificial intelligence, where we’re always looking for, you know, is this something that’s as good as a human? Is it a PhD level or, you know, a grade five level? What level of intelligence are we talking about? And what we miss in that process is that we stop looking at AI as we’ve looked at other technologies.

If you look at the history of technologies, and I take an example of, I take two examples, I’ll take just one right now, but I take two examples of seemingly unintelligent technologies in the book. One is the container, the shipping container, and the other is the barcode. Now the shipping container is very interesting because before, you know, the shipping container came in, we did not have globalization.

We had very local supply chains, we had local manufacturing models, local economies, and arguably the shipping container completely transformed how trade is done. But when the shipping container actually came in as a technology, what its initial effects were read as, you know, a change in how ports operated, because before the shipping container, ports had to handle break bulk cargo, which was not standardized, and so it had to be manually handled. And because the shipping container was a standardized box, it could be loaded and unloaded using cranes.

The containers could be stacked very nicely in ships. And so all of that made ports more automated. The jobs of dock workers went away, yes, so, you know, the shipping container did change, did have an impact at the level of the task, at the level of how cargo was handled.

But if we just focus on that particular point, you know, where the technology was applied, because that’s where we usually focus, how are things changing at the point of application of technology. If we just focus there, the revolution would have stopped at the automation of ports. What happened instead was completely different.

Once ports got automated, the real value of the shipping container came up when trucks, trains, and ships agreed to use the same format, standard format of the shipping container, so that the same container could be loaded and unloaded of different modes of transport. Now, the reason that’s important is that it’s no longer a shipping revolution, it’s a logistics revolution. You had intermodal transport possible.

You could move from any place to any place in the world with the same container. And that essentially dramatically collapsed the cost of moving goods, but also the unreliability associated with moving goods, because things did not have to get stuck at various points in the logistics chain. And when unreliability of transportation went away, suddenly the entire manufacturing and trade landscape changed, because traditionally, manufacturing was handled largely vertically integrated in-house, or it was kept close by.

Your suppliers would be close to you, because you could have the liability in terms of accessing their components. On the retail side, inventory used to be stopped, because there was just unreliability in what would be received. With the shipping container, manufacturing got unbundled, components started being created in different parts of the world.

China rose as a result. I talk about Singapore’s rise, because it bet on this larger shift that was happening, not just on the port. And what ended up happening was that the innovation across industries changed, because now components could compete and component-level innovation could happen, and those components could be recombined to create new products.

And so product quality changed, components changed, the logic of industries changed, global supply chains came into being, and eventually what we see now is globalization happened. So all of this happened when you see all of this, when you go away from the point of application of the technology, which was the port and the train, and see the larger system. And that’s what we’re missing with AI today, and that’s what I argue with the real effects of AI, when you think of the systemic effects.

So Sanjit, that is truly fascinating to me, two elements that you highlight here. The focus is too much on the individual algorithm and not enough in the orchestration of the system and the way everything plays together, which is where most of the benefits or the competitive advantages will be created. Let me give you a parallel with the banking industry.

It’s a bit of a stretch, but we try to bring it back home, and maybe we have then a couple of words about the barcode, which is more into the data economy compared to containers. So in banking in the 1990s, all banks were investing heavily in algorithms for financial engineering. So very complex mathematical formula, and this bank had the best one, that bank has an even better one.

Seems to me like the race of the AI players with the larger language models. There were other banks instead that focused more in terms of how could they build their risk management capability in order to cross all of these different decision-making processes that were otherwise siloed between market risk, counterparty credit risk, because in the end is the way you reshuffle the decision-making process in the organization that makes you more resilient and therefore more competitive facing adversities. So fast forward 10-15 years of this crazy time of financial innovation, the financial crisis started, and all the banks that were priding themselves to have the best-in-class algorithm failed almost miserably.

The others instead, they were capable of creating decision-making without having such an attrition in terms of the consensus of what had to be made in terms of moving forward at the competitive advantage. A bit of a stretch here compared to what you’re saying, but I see the same like many providers of technology focusing on the best-in-class algorithm as if there is a competitive advantage, while in reality is the platform that allows it to govern and orchestrate across the firms or across the ecosystem, the one that ultimately will give you the real advantage, especially thinking that a lot of the algorithms might start becoming small models or open source, changing again, you know, the competition, if you like, and the primacy. Now, the continuous example is very cool because again, it’s not about the latest in technology, the primacy per se, but the way you use it, even more simplified, but how you insert it into a more complex process.

If you think about the barcode, that’s a cool example as well, right? Because a barcode is very simple. It generates data though. You can decide to use the data or not.

So what about Walmart? Yeah, exactly. I think, you know, I wanted to show how different companies can misread the same technology and have vastly different outcomes. And the barcode is one of the best examples of this.

It’s also, as you mentioned, an early example of how data fundamentally changes orchestration and what I call coordination of the entire system. Now, when you think about what happened with the barcode, when barcodes first came in, again, most retailers saw it as a way to improve tasks, right? So you put a barcode on a product, it helps you scan things faster, checkout becomes faster, you can have automated checkout, you don’t need a cashier at the checkout point, the cashier’s job changes, and so on. So most retailers and Kmart, you know, one of the prominent retailers in the US at that time, saw the barcode as a way to automate checkout.

Walmart saw it as a way to redistribute power in the retail ecosystem. Because what Walmart did with the barcode was that it started aggregating all of this information about which products were selling at what rate across its stores. And it started doing that across its entire network of stores, which gave it the ability to create a centralized representation of how inventory was moving across its entire network.

Walmart had already invested in significant IT infrastructure and satellites around that time. And so using all of that, all of these complementary technologies, it created a sort of a new internal system where suppliers had to connect in to see how products were selling. And because Walmart owned all of that data and all of that information, it could exert power on both sides.

It could dictate suppliers on what to stock when. And on the other side, closer to the consumer, it could control what kind of promotions to run for which kinds of products. So because it had these two points of power, it flipped the relationship between the retailer and the supplier.

Because traditionally, suppliers used to guide retailers in terms of what should be stocked and when. The other reason for that was that before the barcode and before Walmart did all of this, retail used to work at a store level. So every individual store manager would work with a supplier.

And the supplier clearly had much more visibility into what was selling across stores than that individual store manager did. And so the negotiation power was very different. But what Walmart did was it centralized its procurement and inventory management, something that is very obvious today, but back then was something fundamentally new.

And that unlocked economies of scale in a chain retail store. And so that’s, again, what flipped the power balance between the two. So a key thing that we see when you think about the Walmart example, it exactly illustrates what we see with data and AI today.

What exactly did the barcode do? The barcode, first of all, made the movement of goods visible. When the movement of goods became visible, the associated decisions around when to stock what and which promotions to run became much better informed. And on the basis of that decision, the right actions could be executed.

And once Walmart, in this case, became the central place which owned that representation of movement of goods, the decision associated with it, and the ability to act on it, everybody had to fit into Walmart’s system and had to be governed by Walmart than the other way around. And so it flipped the power because it created this new visibility. And that’s really what we see with data and now with AI today, as the ability to reorganize power because of the new visibility that it creates.

Sanjit, so you said something that is extremely relevant here, which is the power of shifting power. That is what is happening with AI in terms of how we can allow people to redesign the environment around them and so get control of that through a new coordination perspective. Now, I want to mention here an Italian philosopher that departed a few years ago, Emanuele Severino, who is a giant.

And he has been talking a lot about the potency of techniques, you say technology, and actually the collapse of technology in the attempt just to over-optimize a framework. But one of the key messages that he developed is that a theory, also science, if you like, is not true or false in itself, but starts to be dominant because it’s the one that transforms the world. So if a view does not transform the world, if a technology does not transform the world, it cannot be dominant, it cannot have the real impact of the one you expect.

So now transforming the world starts by, if you like, understanding what the world is, so having a proper representation of the world or a bad representation of the world. So in your book, when you talk about this shifting of power, you define some elements or factors that are needed to create this extra level of coordination that empowers someone to do something that was not possible before. And it starts effectively by representing the world.

So the barcode, in essence, didn’t simply facilitate faster tracking of information, but provided a web of intricate data points that allowed to represent what was happening in better terms. So somebody could basically grab it, you know, as their power, and reorganize an entire economy, or if not more closely, an entire firm around that. So which are these coordination factors that allow AI to restructure power without attaching the tasker? Yeah, I think this is something which is quite interesting, where, you know, traditionally, a lot of coordination involved some form of consensus between different parties, right? So the shipping container, for its value to be unlocked, there needed to be consensus between ships, trucks, and trains, that they would use the same format of the container, the same contract, and so on.

So that’s what I call coordination through consensus. Once you have consensus, it becomes easy to coordinate. One of the key things that AI does is it allows coordination without consensus, right? And you might, you know, you might think that a lot of coordination today actually happens without consensus, but a lot of that happens in consumer markets, where companies have huge power over how the market is organized.

So think of, you know, Amazon coordinating sellers with buyers and matching the right seller with the right buyer, it’s able to do that, because it’s able to extract data from you at an extraordinary rate, and use that market power to then force the suppliers to agree to whatever, you know, rules of governance data are. But you can’t enforce that in more complex industries, where there are multiple sources of power. So banking is a classic example, healthcare, construction, industrial sectors, there’s no one single player that can, you know, capture data across the industry and coordinate on that basis.

There’s no existing consensus in a lot of cases where players already want to work together. So in these cases, where in the absence of consensus, and in the absence of dominant players, AI provides a really powerful mechanism to drive coordination without consensus. And that coordination without consensus happens by first solving a consumer problem, or solving a stakeholder problem more broadly, of managing fragmented information and making decisions based on fragmented information.

I’ll take two examples, which will bring this to life. So if, you know, we’re just back, all of us are back from our summer travels, when you’re traveling, you realize that travel planning and execution is a fragmented problem, because your flights are booked on one side, your hotels booked on another side, and your activities are planned somewhere else, there’s something in spreadsheets, something on Google Maps in your favorites list. And so it’s, it’s very scattered.

But AI provides the ability to take all of this fragmented information and make sense of it. One of the key things that I did, you know, was just talk to Chad GPD on during my vacation and just say, this is where I am today, this is what I want to do, what I don’t want to do, give me a full itinerary, which fits these constraints. That is something that would not have happened in the past.

Now, if you overlay that on top of the ability of an AI assistant to also go and negotiate those exact activities from the right players, you could have somebody working entirely on your behalf and manage that coordination without consensus. So the ability to bring fragmented information together, you know, you throw in your, your ideas, your ticket PDFs, your Google spreadsheet, and the ability to take all of that fragmented information and manage your itinerary on that basis is what I think of as coordination without consensus. It’s a unique property that previous technologies have not given us just because of the inability to handle unstructured information.

Today, AI is at a place where unstructured information can be managed and a structured representation can be created around it, with which multiple players can coordinate their activities and decisions. So Sanjeev, first of all, let me tell you that Chad GPD had an easier time with you because I know you had vacation in Italy, and even a wrong route in Italy is a wonderful route. So we don’t have the time to suggest, you know, a way to go.

But this is an important aspect that you’re highlighting is about implicit knowledge that becomes explicit knowledge. And once you allow you to map and represent the knowledge that is used inside an unformatted data-driven economy, then you can get a power in order to basically reorchestrate the way people interact around that piece of information. It reminds me of some conversation I’ve been having recently with chief information officers of primary banks worldwide about how to use artificial intelligence to modernize core banking.

There is definitely a possibility now to automate the coding and accelerate that, like translating, let’s say, a monolithic set of programs into something more modular and flexible. But the key element that some are already identifying is not that process of automation, which would be a task. The issue is somewhere else.

And it lays in the fact that what technical depth is, to me, is not the server used for a process, but it’s the fact that in monolithic codes, many programmers embedded the business logic needed to run a complex process. And, you know, that goes through layers and layers of codes, years after years, where the programmer in the bank lost some of the causality of the system. And now that programmer leaves because the guy retires, it brings with itself not just the knowledge of code, but the knowledge of the implicit knowledge in the system.

And then with AI, what you can do is to understand that code and extract the rules and the understanding of what happens in order to replace those rules somewhere else. But you need to be paying attention here because you are asking now coders with a higher capability, think about those that create a orchestration to re-embed the business logic into the orchestration layer of AI with the risk of recreating implicit knowledge. But those, to me, that instead understand these aspects and grab the knowledge to keep it explicit to the understanding of everybody else, we will be able to interact with that.

By keeping it explicit, they will force everybody around the process to coordinate themselves differently in the way they keep on creating and transforming the core of the financial institutions. So now I found the parallel in what you’re saying, in thinking about what the CIOs are thinking for grabbing an asymmetrical competitive advantage by applying not just the effort task, but AI for the orchestration of knowledge inside the organization. So Sanjay, what is the key message or the piece that you believe is resonating the most about your book that you want to share with the Bankers Bookshelf audience today? Yeah, I think the key message that has really resonated with people is that very often when we talk about AI, we sort of go into two extremes.

We think about AI changing everything and saving our world and solving all our problems and creating incredible prosperity and everybody will then go into universal basic income. And on the other hand, we talk about everybody losing their jobs and AI taking away the jobs. And we have these two extreme views, techno-optimism and complete doom and gloom.

And neither of these views is actually completely correct. There’s a lot of nuance that they’re missing. And the point that the book tries to make is that it’s very likely and it’s actually most likely for the entire pie to increase and pie to grow and yet not be shared equally.

And it’s not something new. We’re seeing this for the last 30, 40 years, first with financialization, then with the rise of the platform economy, that the pie dramatically grows, but it’s not necessarily divided equally. And what the book talks about is it lays out a very clear structure and a clear set of systems design principles that determine how the pie grows through what I call the unbundling and the rebundling of traditional systems and how it is then divided to the new tensions that emerge in the system.

So the entire book talks about these with a lot of stories, a lot of examples, both historically, a lot from sports, from F1, from football, and so on. But then it brings it back to what’s happening with AI and really talks about why these tensions come up when new systems emerge around new technologies and how to resolve these tensions. And so I think that’s been a topic that’s not been discussed enough and that’s what’s really caught on with the readers quite well.

So Sanjay, it’s something that I want to ask you to conclude this episode. You actually suggest people that ask themselves, what do I have to do to promote my career in the future? Something relevant to say, to look at the next task, but look at where the constraint is removed. Can you explain? Yeah, I think one of the key things that we need to understand is that our jobs are not simply a set of tasks.

Even though we say our job is to be a teacher or a painter or the plumber, and those names are associated with tasks, teaching, painting, plumbing, but the jobs are not just about tasks. And a really good example of this is to really think about what happened with typists when the word processor came in. Because you would think that the job of a typist is to type, and you would not be wrong.

A lot of what they did was typing. But when the word processor came in, a lot of people would have told them that word processors won’t take your job, but typists who learn word processors will take your job. And so if you learn word processors, you’ll be fine.

But what ended up happening was that the job of a typist completely went away. In fact, instead, everybody started typing. So it’s not like technology took away the task.

The task is still a human task, but the job no longer exists. And the reason for that is that the job was never about typing. The job was about managing the constraint in the old system.

Before the word processor, the constraint associated with typing was editing. Editing was expensive. If you made others, fixing that was expensive.

So the constraint enabled the job of the typist to exist. When the word processor came in, the constraint was removed, and so that job did not need to exist anymore. So the point that I make in reshuffle with the typing example and the other examples is look for the constraint and see where it’s moving.

Because typically in today’s economy, there are three constraints or three categories of constraints around which jobs are structured. They’re constraints of scarcity. So there could be scarcity of certain skill because of where you’re located or how difficult it is to access that skill or to train on it.

There could be constraint associated with managing risk, and there could be constraints associated with managing coordination. And one of the things that we see is that with AI coming in, a lot of the scarcity associated with skill might actually go away. But there would be constraints associated with risk and coordination that you should constantly look for.

And a very simple example to think of this is when you think of the job of an anesthesiologist who administers anesthesia when you’re in an operation or a procedure, most of the tasks involved in their job have already been automated. Machines are doing it. And yet anesthesiologists are paid very well.

And the reason they’re paid very well is not to perform those tasks, but to manage the risk in a very high stakes setting. And so we need to understand that even if AI comes in and takes away specific tasks, chips away at those tasks, looking for new tasks to do and new skills to learn is not enough. Look for where the new system is moving.

What are the new risks? What are the coordination breakdowns that are happening? What are these new constraints that are emerging? And try to understand how can you own that constraint and then put your skills around it. So skilling is important, but not unless you have a clear view of what the constraints are. So Sanjit, you may explain what people smell, but I don’t yet understand in full synthesis.

We ran a survey on 3,000 executives worldwide two years ago, a year and a half, about the skills required to the workforce going forward. And there was a list of skills. The top one in 2022 was STEMs.

Okay. So high engineering. And of course, out of a list of 14, creativity was down because they claimed that engineers are not creative or economists.

I’m an economist, I’m creative, but that’s the rule of thumb. And there were a lot of elements in between like teamwork ability and time management and stuff. And two years ago, when we asked the same question, proficiency in STEM went to the bottom.

Creativity didn’t go up, but what went to the number one was time management and teamwork ability. Because people realize that there’s so much complexity in the system. What they need is to re-coordinate the way we work.

So implicitly they were expressing the need to succeed, to coordinate themselves differently. And that’s what I can do them without thinking about individual tasks. Now, going back to your typist example, I confess to the audience that I tap with two fingers, left and right hand.

And I know that this will be disintermediated as well with voice, but likely that constraint will still be down through my voice. So I remain relevant in a voice-driven typing world. So Sanjeev, thanks very much for this conversation.

Where can the audience follow your work and then reach out to you if they need? Yeah, absolutely. You can grab a copy of Reshuffle on Amazon. It’s available there in hardcover, paperback, audio, kindle formats.

You can follow my newsletter. It’s called platforms.substack.com, where I publish every week on this topic. And you can learn more about my work at www.platformthinkinglabs.com. So platforms.substack.com and then Amazon to get your copy of Reshuffle.

Reshuffle is my reading recommendation this month to learn about a framework to understand and master how AI can shift the power across firms and entire ecosystems. Thanks for joining, Sanjeev. Thank you, Pablo.

And I invite all of you to stay tuned for more conversations with book authors and researchers that shape or increase our understanding of the banking present and its fintech future. Don’t forget to subscribe to the Bankers Bookshelf to receive your polite notification about new episodes on your preferred platform. We stream on Spotify, Apple Podcasts, Amazon, and YouTube.