I still hear owners say, "We're not ready for an AI strategy." Then I ask what they're running today. Someone has ChatGPT open. Marketing bought a writing tool. Ops is testing a chatbot on the contact page. That is a strategy — just not one anyone wrote down, funded correctly, or owns.
Every CEO needs an AI strategy. Not because the consultants said so. Because without one, your company will accumulate tools, rent intelligence, and call it progress while a competitor builds the layer that compounds. I learned that rebuilding CI Web Group from 320 people to 38 — not as a layoff story, as an architecture story — with more than ten million dollars invested in the AI layer that replaced coordination headcount.
What an AI strategy actually is
It is not a list of vendors. It is four decisions only the CEO can make:
- What layer are we building? Chat is step one on the maturity ladder. Strategy names where you are going — infrastructure, intelligence, agents — not where you demoed last Tuesday.
- Who owns it? If the answer is "IT" or "the agency," you don't have a strategy. You have outsourcing. Read what the CEO job becomes when the company runs on systems.
- What compounds? Bolt-on tools evaporate context. Built AI turns every interaction into institutional memory. Strategy chooses compound interest over monthly rent.
- What won't we automate? That's the soul half of systems and soul — the judgment you refuse to delegate.
What "we'll wait" actually costs
Waiting feels prudent. It isn't. It means your data stays trapped, your team learns the wrong lesson (AI is not ChatGPT), and the gap widens quietly. I've been in rooms since 2006 building CI Web Group, and since 2008 when I made the Daikin bet that infrastructure would matter more than headcount. The owners who waited didn't lose to a smarter model. They lost to a better operating system someone else installed while they were "evaluating."
Strategy is how you stop evaluating and start building — with a dated scoreboard, not a vibe.
The SME inversion applies to the corner office
Most CEOs still think their job is to be the answer key. The SME inversion says the opposite: engineer your replacement for repeatable decisions so you can stay in the work that isn't repeatable — culture, ethics, customer truth, the call that keeps a company human.
That is strategy in practice. Not a deck. A workflow map that shows what agents carry, what humans carry, and what you personally will never hand off.
Strategy is not a roadmap you hide
The worst AI strategies live in a drawer — written once for the board, never referenced again, never tied to spend, never tied to hiring, never tied to what you refuse to automate. Real strategy shows up in the budget. It shows up in who you promote. It shows up in what you kill when a vendor demo doesn't survive the soul test.
When we rebuilt around agents — the customer hand, the operator hand, the market hand I describe in Three Hands — strategy stopped being a slide and became a workflow map. That's the difference between CEOs who "have an AI initiative" and CEOs who have a company that compounds.
What I tell rooms full of owners
On trade stages I say the same thing I say in boardrooms: you don't need to predict every model. You need to own the layer. You don't need forty tools. You need one architecture decision and the courage to rebuild when the scaffold stops teaching you. I made that call with Daikin in 2008 when the comfortable choice was wait. Strategy is that kind of call — dated, owned, public enough that you can't quietly drift back to average.
Strategy also names what you won't do — the shortcuts that demo well and compound badly. No second AI vendor without integration. No public "AI company" claims until the layer remembers. No automation that fails the soul test. Constraints are how strategy becomes real.
Action
Block ninety minutes. Write one page: current state on the ladder, target state in twelve months, owner name, budget line, one metric with a review date. If your leadership team needs the argument in a room, book the keynote. If you want the longer letter about conviction under pressure, read Hands Up. Strategy that stays in your head isn't strategy. It's hope.