Brief

The Pragmatic CEO’s Guide to AI

Although no one can perfectly predict what AI success will look like in their industry, the best CEOs are focused on pushing AI to failure … and beyond.

By Dunigan O'Keeffe, Jamie Cleghorn, Raj Pherwani, and Martin Toner

Published on 十一月 18, 2025
min read


When we look at the amount of money pouring into artificial intelligence, we can’t help but think of the movie Brewster’s Millions, where Richard Pryor plays a man who must spend $30 million in 30 days in order to inherit a much larger sum. That’s not to imply that companies are deliberately trying to blow as much money as they can as quickly as they can; it simply reminds us how hard it is to spend large amounts of money both quickly and well. Over the last five years, Alphabet, Amazon, Meta, and Microsoft have invested a combined total of nearly $1 trillion in capex, most of it focused on AI. It is inevitable that some of that money will be wasted, but here’s the thing: can you tell which of these dollars are being invested and which are being wasted? Not from this vantage point you can’t. Other investors in AI right now are in more or less the same position. Our goal here is not to make bold predictions about the future—to do so is inevitably to give a hostage to fortune. Instead, we want to share what we’ve learned from our conversations with C-suite executives and board members at the tech companies that are at the bleeding edge of creating AI-driven solutions and at the companies trying to lead their industries in the adoption of these new technologies. Here are the three things that CEOs and C-suite executives need to think about now. And we mean, right now. 1. It is the CEO’s responsibility to make the organization fit for the future; that which cannot be avoided should not be delayed Conversations about what AI means for business are close to meaningless. The question is: What does AI mean for your business? There are many definitions of strategy, with a serviceable one being “the art and science of allocating scarce resources in order to create sustained competitive advantage.” This provides a simplifying lens through which to view the kaleidoscope of opportunities and threats presented by AI: Where can this technology help me build sustained competitive advantage and where is it a threat to the advantages I have today? A simplifying lens through which to view AI: Where can it help me build sustained competitive advantage? What is the basis of competition in your industry? How much of it is physical vs. digital? To what extent are your profit pools insulated (for now) by the need to make and move physical objects from place to place? If your business primarily competes on the basis of gathering, consolidating, synthesizing, analyzing, and acting on data, your world is going to change fast. Trust us, that’s basically what a consulting firm does, so we understand the need not just to embrace change but also to lead it. “The future is already here—it’s just not evenly distributed,” said the science fiction author William Gibson. Well, AI is here, and it is impacting different sectors at different rates. Within each sector, the winners will be determined by who is willing to act bold enough, fast enough. 2. If you’re still experimenting with use cases, you are behind; the value comes from reinventing how work actually gets done Everything has to start somewhere, and for AI that has been use cases. This is natural because use cases are observable, discrete, and tangible. But they are also too small to create the kind of value that might justify $1 trillion in capex. Perhaps nothing can justify that amount, but if anything will, it will be using AI to comprehensively redesign the processes that create the most value in your business. That is easier said than done. Businesses are living organisms, made up of autonomous human beings. The most important processes are often informal and unwritten. AI’s ability to parse massive volumes of unstructured data is a major leg up. It presents an opportunity to compare that current, messy state to how you might design the business if you could start with a clean sheet of paper. What would your business look like if you redesigned it from scratch to be an AI-native enterprise? How different would your business processes be if you had a one-person HR team? Would you like to find out? What would your business look like if you redesigned it from scratch to be an AI-native enterprise? 3. You don’t need a roadmap; you need to get started. Identify the next few stepping stones and move with conviction This may be out of character coming from a group of strategy consultants, but this may be one case where capital “S” strategy is not the right place to start. In fact, we heard this loud and clear in our recent discussions with executives working at the coalface of implementing AI. The chief strategy officer of a Fortune 100 company put it this way: “The strategy questions are massively important, but until we can start demonstrating ROI through our current AI initiatives, it will be hard to get anyone to pay attention to strategy.” Fortunately, although many companies have been frustrated by the lack of early returns, we are seeing examples of companies deploying AI with real financial results. The best thing companies can do is to start deploying AI in ways that meaningfully change their business. If you try to figure out not just the next couple of moves but the whole chessboard, you will find yourself outmaneuvered: The game is moving too quickly. The situation cries out for launching a portfolio of micro-battles to accelerate learning and scale what works. If you try to figure out the whole chessboard, you will find yourself outmaneuvered: The game is moving too quickly. Micro-battles are not pilots. Just as individual use cases don’t deliver meaningful change, pilots are typically set up for success, tackling a small problem with big resources. To conduct a proper micro-battle, you need to pick a really tough business process problem with a lot of value at stake and figure out how to solve it. Micro-battles are designed to run hard at the first potential point of failure and either solve it and move to the next one or go back to the drawing board. In the time it would take you to think a few more moves down the chessboard, you could already be out there testing and learning, scaling what works and sharing lessons learned about what doesn’t. The “micro” in micro-battle is not meant to imply small, per se, but rather contained, circumscribed, manageable. Think big enough to make a difference. Shaping your AI future starts with these questions Anybody who tells you they have all the right answers on AI is lying. At least we have the right questions. Our conversations with executives and board members grappling with this topic have helped us to distill these critical external and internal questions that every C-suite needs to come to grips with. Gameboard shifts: How will AI structurally change our market? Will the value chain reorganize? Are profit pools moving with it? What assets and capabilities will win in the future? How are customers, behaviors, and cost profiles shifting? Strategic signals: How fast do we have to move? Are new-to-category customers going to insurgents? Are customers splitting their wallets? Are the highest net-present-value talent making their career bets in new places? Are there marked new capital flows going on in and around your profit pools (both institutional capital and M&A)? Technology disruption assessment: What will be required for us to win? What is the AI exposure of our cost structure and value proposition? Where is our industry on the AI experience curve and which company is leading the way? Is the gravity of the most valuable data—today and tomorrow—shifting? What level of technology investment is required to create advantage? What is the ratio of retiring technology debt to investing in new capabilities? Transformation agenda: How will we change our business? What is your overall ambition? How will you deliver against the raw customer need? How will you build the capabilities and assets you need to win? Do you have the talent, culture, and operating model to scale faster than the competition? It is natural for any of us—linear, logical folks that we tend to be—to want to work through these questions from top to bottom, but the world is seldom that tidy. We see every question on the list popping up in our conversations with senior executives and board members. But for all of them, the first step is to understand where their industry is in the AI penetration curve and what opportunities there are to create sustained advantage (or to avoid the erosion of previous advantages). After that, the most important questions are, “How do we get started?” and “How fast can we go?”

When we look at the amount of money pouring into artificial intelligence, we can’t help but think of the movie Brewster’s Millions, where Richard Pryor plays a man who must spend $30 million in 30 days in order to inherit a much larger sum.

That’s not to imply that companies are deliberately trying to blow as much money as they can as quickly as they can; it simply reminds us how hard it is to spend large amounts of money both quickly and well.

Over the last five years, Alphabet, Amazon, Meta, and Microsoft have invested a combined total of nearly $1 trillion in capex, most of it focused on AI. It is inevitable that some of that money will be wasted, but here’s the thing: can you tell which of these dollars are being invested and which are being wasted? Not from this vantage point you can’t.

Other investors in AI right now are in more or less the same position.

Our goal here is not to make bold predictions about the future—to do so is inevitably to give a hostage to fortune. Instead, we want to share what we’ve learned from our conversations with C-suite executives and board members at the tech companies that are at the bleeding edge of creating AI-driven solutions and at the companies trying to lead their industries in the adoption of these new technologies.

Here are the three things that CEOs and C-suite executives need to think about now. And we mean, right now.

1. It is the CEO’s responsibility to make the organization fit for the future; that which cannot be avoided should not be delayed

Conversations about what AI means for business are close to meaningless. The question is: What does AI mean for your business?

There are many definitions of strategy, with a serviceable one being “the art and science of allocating scarce resources in order to create sustained competitive advantage.”

This provides a simplifying lens through which to view the kaleidoscope of opportunities and threats presented by AI: Where can this technology help me build sustained competitive advantage and where is it a threat to the advantages I have today?

A simplifying lens through which to view AI: Where can it help me build sustained competitive advantage?

What is the basis of competition in your industry? How much of it is physical vs. digital? To what extent are your profit pools insulated (for now) by the need to make and move physical objects from place to place?

If your business primarily competes on the basis of gathering, consolidating, synthesizing, analyzing, and acting on data, your world is going to change fast. Trust us, that’s basically what a consulting firm does, so we understand the need not just to embrace change but also to lead it.

“The future is already here—it’s just not evenly distributed,” said the science fiction author William Gibson. Well, AI is here, and it is impacting different sectors at different rates. Within each sector, the winners will be determined by who is willing to act bold enough, fast enough.

2. If you’re still experimenting with use cases, you are behind; the value comes from reinventing how work actually gets done

Everything has to start somewhere, and for AI that has been use cases. This is natural because use cases are observable, discrete, and tangible. But they are also too small to create the kind of value that might justify $1 trillion in capex.

Perhaps nothing can justify that amount, but if anything will, it will be using AI to comprehensively redesign the processes that create the most value in your business. That is easier said than done. Businesses are living organisms, made up of autonomous human beings. The most important processes are often informal and unwritten. AI’s ability to parse massive volumes of unstructured data is a major leg up. It presents an opportunity to compare that current, messy state to how you might design the business if you could start with a clean sheet of paper.

What would your business look like if you redesigned it from scratch to be an AI-native enterprise? How different would your business processes be if you had a one-person HR team? Would you like to find out?

What would your business look like if you redesigned it from scratch to be an AI-native enterprise?

3. You don’t need a roadmap; you need to get started. Identify the next few stepping stones and move with conviction

This may be out of character coming from a group of strategy consultants, but this may be one case where capital “S” strategy is not the right place to start.

In fact, we heard this loud and clear in our recent discussions with executives working at the coalface of implementing AI. The chief strategy officer of a Fortune 100 company put it this way: “The strategy questions are massively important, but until we can start demonstrating ROI through our current AI initiatives, it will be hard to get anyone to pay attention to strategy.”

Fortunately, although many companies have been frustrated by the lack of early returns, we are seeing examples of companies deploying AI with real financial results. The best thing companies can do is to start deploying AI in ways that meaningfully change their business.

If you try to figure out not just the next couple of moves but the whole chessboard, you will find yourself outmaneuvered: The game is moving too quickly. The situation cries out for launching a portfolio of micro-battles to accelerate learning and scale what works.

If you try to figure out the whole chessboard, you will find yourself outmaneuvered: The game is moving too quickly.

Micro-battles are not pilots. Just as individual use cases don’t deliver meaningful change, pilots are typically set up for success, tackling a small problem with big resources. To conduct a proper micro-battle, you need to pick a really tough business process problem with a lot of value at stake and figure out how to solve it. Micro-battles are designed to run hard at the first potential point of failure and either solve it and move to the next one or go back to the drawing board.

In the time it would take you to think a few more moves down the chessboard, you could already be out there testing and learning, scaling what works and sharing lessons learned about what doesn’t. The “micro” in micro-battle is not meant to imply small, per se, but rather contained, circumscribed, manageable. Think big enough to make a difference.

Shaping your AI future starts with these questions

Anybody who tells you they have all the right answers on AI is lying. At least we have the right questions. Our conversations with executives and board members grappling with this topic have helped us to distill these critical external and internal questions that every C-suite needs to come to grips with.

Gameboard shifts: How will AI structurally change our market?

Will the value chain reorganize? Are profit pools moving with it?
What assets and capabilities will win in the future?
How are customers, behaviors, and cost profiles shifting?

Strategic signals: How fast do we have to move?

Are new-to-category customers going to insurgents? Are customers splitting their wallets?
Are the highest net-present-value talent making their career bets in new places?
Are there marked new capital flows going on in and around your profit pools (both institutional capital and M&A)?

Technology disruption assessment: What will be required for us to win?

What is the AI exposure of our cost structure and value proposition? Where is our industry on the AI experience curve and which company is leading the way?
Is the gravity of the most valuable data—today and tomorrow—shifting?
What level of technology investment is required to create advantage? What is the ratio of retiring technology debt to investing in new capabilities?

Transformation agenda: How will we change our business?

What is your overall ambition? How will you deliver against the raw customer need?
How will you build the capabilities and assets you need to win?
Do you have the talent, culture, and operating model to scale faster than the competition?

It is natural for any of us—linear, logical folks that we tend to be—to want to work through these questions from top to bottom, but the world is seldom that tidy. We see every question on the list popping up in our conversations with senior executives and board members.

But for all of them, the first step is to understand where their industry is in the AI penetration curve and what opportunities there are to create sustained advantage (or to avoid the erosion of previous advantages).

After that, the most important questions are, “How do we get started?” and “How fast can we go?”