AI Strategy for CEOs: Own It, Do Not Delegate It

By Andreas Pettersson, founder of Leaders ADAPT and a former Canon AI executive who built and sold an AI company before ChatGPT existed.

Let me be direct. Most CEOs I talk to have already delegated their AI strategy. They just will not say it out loud.

They handed it to IT. Or to a consultant. Then they waited for a plan to come back. And what came back looked responsible. A vendor shortlist. A governance policy. A pilot running somewhere safe. Everyone felt good about it.

Here’s the problem. You don’t have an AI strategy. You have a containment plan. And those are opposites.

I have stood on both sides of this. I scaled real machine learning and computer vision at Arcules years before ChatGPT, then watched the market commoditize the exact thing we had built from scratch. Here is what that taught me. The advantage was never the technology. The technology gets cheap. The advantage was judgment about where to point it. That part never commoditizes, and it is the one part of AI strategy for CEOs nobody can hand you.

Why is AI strategy a CEO job and not an IT job?

Because AI is not a technology decision. It’s a decision about where to multiply force in your business.

That sounds like a slogan. It isn’t. It is the whole thing. Think about what AI actually is at your level. It is a force multiplier. One unit of force in, ten units out. A lever. And like any lever, it amplifies whatever it touches. Point it at a strong process and you get a much stronger one. Point it at a broken one and you scale the brokenness. Point it at the wrong corner of the business and you get a faster version of nothing.

So the real work is not technical. The real work is placement. Where do you put the lever? And just as important, what do you choose not to multiply yet? That is a CEO decision. It cuts across functions. It touches your margins, your people, your competitive position. No IT department can make that call, because none is accountable for the whole business. You are.

Now, here is what happens when you delegate it anyway. You hand AI to a function that is brilliant at its actual job. IT is built to secure, integrate, govern, and contain. So that is what it does. And containment is the exact opposite of multiplication.

The trap is not that IT fails. The trap is that IT succeeds at exactly what it was designed to do, and ruins AI for you as a force multiplier.

Read that twice. It is the part almost everyone misses. You are not waiting on a failure. You are getting a success at the wrong objective. Twelve months later you have an approved vendor list, a policy document, and a pilot in accounts payable. You are technically “using AI.” Nothing has multiplied. No margin change. No speed change. Nothing you can explain when the board follows up.

You don’t have a technology problem. You have a leadership problem. And the data agrees. RAND found that more than 80% of AI projects fail, about twice the rate of regular IT projects. S&P Global reported that the share of companies abandoning most of their AI initiatives jumped from 17% in 2024 to 42% in 2025. Those are not model failures. Those are placement failures.

The IT delegation trap, and what it actually costs

Let me tell you the story that makes this concrete. It is not an AI story. It is older than that.

In the late 1990s, two companies looked at the internet. One was Borders. Over a thousand bookstores. Decades of history. Real scale. The other was Amazon, a startup that did not exist yet at that scale.

Borders treated the internet as a technology to manage. So they did the responsible thing. They delegated their online strategy. They ran pilots. They waited to understand it before they committed. They asked the safe questions, the ones a careful operator asks. Vendor security. Integration timelines. Return on investment. And in 2001 they made the decision that looked the most reasonable of all. They outsourced their entire online business to Amazon. Why build it when someone else can run it better?

Within a decade, Borders was gone.

Here’s the thing. Borders did not fail because they were careless. They failed because they were careful. They were prudent. They were disciplined. They applied all of that discipline to the wrong frame. They asked, how do we manage this technology safely? Amazon’s leadership asked a completely different question. Where does speed compound? Where does reach multiply? Where does learning accelerate?

Same internet. Same era. Same technology available to everyone. Different placement. Different outcome.

That is the IT delegation trap in one example. Borders handed a force multiplier to the part of the org that manages technology, and that part managed it, beautifully, right out of relevance. Reasons aren’t results. Borders had every reason. Amazon got the result. You are making the Borders decision or the Amazon decision right now, and the Borders one feels safer every time. That is exactly why most leaders get it wrong.

Stop asking for an “AI strategy.” Demand a leverage map.

So what do you do instead? You change what you ask for.

The next time someone hands you an “AI strategy,” push it back. You are not asking for an AI strategy. You are asking for a leverage map. Not a vendor list. Not a policy. Not a slide about adoption metrics. A map of where multiplying force would actually change how the business operates. Not where can we use AI. Not what is easy to automate. Not which vendor has the best demo. Those are the wrong questions, and they produce policy documents instead of leverage.

The right question, the one that opens the whole thing, is this.

Where does leverage actually compound here?

Ask that first. Always. Before anyone mentions a tool, a model, or a vendor. It does most of the work, because it forces the conversation onto outcomes instead of activity. A bottleneck that, if you broke it, would speed up everything downstream. A decision your team makes a hundred times a week where being faster and sharper would compound across the whole system. That is leverage. That is where the lever goes.

Then ask the question almost nobody asks. What must not be multiplied yet? Where would pointing AI right now actually make things worse, given your culture, your trust levels, your incentives as they are today? Because a force multiplier amplifies weakness just as fast as strength. Restraint is not caution. Restraint is the strategy. The CEOs who win are not the fastest. They are the most deliberate about what they leave untouched.

When you have those two answers, you have something a containment plan can never give you. You have a leverage map. And it is yours. It cannot be outsourced, because only you can see the whole board. Every serious conversation about AI strategy for CEOs ends up here, at placement and restraint, decided at the top.

How is this different from AI strategy consulting for CEOs?

Fair question. Once leaders realize this is their job, the instinct is to hire it out. Get AI strategy consulting for CEOs and senior teams. Let the firm tell you where the leverage is.

The difference between AI strategy for CEOs and AI strategy consulting is simple: who holds the map when the engagement ends.

Be careful here. Use a consultant the way you would use a good analyst, to inform your judgment, never to replace it. The moment you let an outside party define your AI problem, you stop being the captain of your ship and become cargo on someone else’s route. They frame the problem in terms of what they can sell or staff. That is not malice. That is gravity. Every firm bends your strategy toward its own offering.

The placement judgment cannot be delegated, because the accountability cannot be delegated. A consultant walks away after the engagement. You live with the result for years. So bring in expertise for the parts that are genuinely technical, and keep the leverage map in your own hands. You can rent the building. You cannot rent the deciding.

What owning AI strategy for CEOs looks like this quarter

You do not need a transformation program. You need to take the map back.

Sit your leadership team down. Do not say the word AI. Ask three questions. Where would multiplying force change our outcomes? What must we not multiply yet? What decision would this let us revisit faster? Write down what you hear. That is the start of your leverage map, built by you, not a vendor.

Then pick one point on that map, the single place where breaking a bottleneck would compound the most. Name the outcome in plain words with a number attached. Put one person in charge of it, not a committee. Give it a few weeks, not a roadmap. Then look at the result honestly and decide to expand it or kill it.

That is the whole motion. One leverage point. One owner. One short loop. One honest review. Do that, and you are leading AI instead of containing it.

Expect the first loop to be humbling. The bottleneck you were sure about turns out to be a symptom, the real one sits two steps upstream, and the owner you picked finds it in week one. That is not failure. That is the map correcting itself, which is what a living strategy is supposed to do. By the third loop, your team stops bringing you tools and starts bringing you leverage points, and that shift tells you the ownership took.

The CEOs who win at AI are not the most technical people in the room. They are the ones who refuse to delegate the thinking. Decide first. The rest follows.

Where to go next

For the full picture of leading AI without being technical, start with the hub, AI for CEOs. For the practical playbook of running it week to week, read The AI Playbook for CEOs.

If you want to make these decisions in a room of CEOs facing the same calls, that is what the AI Executive Mastermind is built for. And if you want the IT delegation trap, the leverage map, and the full framework in one place, it is all in the book, AI Leadership Mastermind.

Your one move this week: stop waiting for an AI strategy to come back from someone else. Ask your team where leverage actually compounds, and start drawing the map yourself.

Join the AI Executive Mastermind | Get the book

Frequently asked questions about AI strategy for CEOs

Is AI strategy a CEO responsibility or an IT responsibility?

It is a CEO responsibility. AI is not a technology decision, it is a decision about where to multiply force in the business, and that cuts across functions, margins, people, and competitive position. IT is built to secure, integrate, and govern technology, which is essential but the opposite of multiplying it. Delegate AI strategy to IT and you do not get a strategy. You get a containment plan.

What is a leverage map?

A leverage map is what you should demand instead of an “AI strategy.” It is a picture of where multiplying force would actually change how your business operates: which bottlenecks, if broken, would speed up everything downstream, and which frequent decisions would compound if they got faster and sharper. It starts with one question, where does leverage actually compound here, asked before anyone names a tool or a vendor.

Why do most AI initiatives fail?

Not because the technology is weak. RAND found more than 80% of AI projects fail, about twice the rate of regular IT projects, and S&P Global reported companies abandoning most of their AI initiatives rose from 17% in 2024 to 42% in 2025. The AI failure studies executives keep quoting all say the same thing: most initiatives never turn a profit, while a handful see rapid revenue and profit acceleration with the same tools. The gap is leadership and judgment about placement, not the technology.

Should a CEO hire AI strategy consultants?

Use them to inform your judgment, not to replace it. The moment you let an outside firm define your AI problem, you become cargo on someone else’s route, because every firm bends the strategy toward what it can sell or staff. Bring in outside expertise for the genuinely technical parts, and keep the placement judgment, the leverage map itself, in your own hands. You are the one accountable for the result years from now.

How much time should a CEO personally spend on AI strategy?

Less than most expect, and more regularly than most deliver. AI strategy for CEOs is measured in decisions, not hours: one working session with your leadership team to draw the first leverage map, then a short weekly loop on a single leverage point with a single owner. A few focused hours a month keeps the map in your hands. The moment it turns into a standing committee, you have drifted back into containment.

What is the IT delegation trap in AI strategy?

The IT delegation trap is handing AI strategy to the IT department because it feels technical. IT can run the infrastructure and tools, but it cannot decide where AI should multiply the business or what to leave alone. Those are leadership calls. Delegate the plumbing, keep the strategy, or AI drifts into busywork.

Sources

• MIT, The GenAI Divide: State of AI in Business 2025
• RAND, The Root Causes of Failure for AI Projects
• S&P Global Market Intelligence, Generative AI shows rapid growth but yields mixed results