3Idea 3 of 5
An AI strategy will get you fired.
Strategy is where a company will make money and why it will win. AI is a capability. A capability relative to a strategy the company has already chosen. Put the capability first and the board will be right to be nervous.
Every board is asking for it. Every vendor is selling it. So most companies now have a document called an AI strategy, and inside it are pilots, a shortlist and a budget. What is missing is the only thing that matters: which product-market the capability is for, and which line of that product-market's P&L it is meant to move.
This is not a new mistake. It is the same mistake companies made with digital, and before that with quality, and before that with IT. A capability elevated to a strategy produces spending without a number attached. Eighteen months later the board asks what it bought, and the honest answer is a vendor relationship.
What AI actually changes
Analysis. That is the whole of it and it is enormous. Pricing intelligence in an aftermarket business where margin has leaked for three years. Demand sensing in an OEM line exposed to a tariff review. Churn prediction in the managed-services book that sets a PE-backed company's exit multiple. Each of these is a capability for a product-market that has already been ranked and chosen. Each has an owner and a P&L line. The vendor list follows from that; it does not lead.
What it does not change
Which businesses the company is in. Which it should exit. Who owns the number. What the board will accept. These are decisions, and no model makes them. A company that has not done the strategy work will use AI to analyse the wrong things faster. A company that has done it will find that AI makes every product-market owner a better strategist than the head office used to be.
So the answer to the board is not an AI strategy. It is a strategy, with two or three AI capabilities inside it, each tied to a number someone has signed for. That answer takes a page. It also keeps the CEO's job.
In action: AI applied to product-markets, not to the company
Three illustrative companies, three answers to the same board question. In every case, two capabilities, each tied to a product-market and a P&L line.
Industrial, $600M
Pricing intelligence in Aftermarket, where margin has leaked for three years. Demand sensing in OEM ahead of the tariff review.
Franchisor, 380 units
Local demand and pricing for the delivery-heavy urban units. Labour scheduling across the Sunbelt group. Both at the address level.
PE-backed, year two
Churn prediction in managed services. Renewal pricing in maintenance contracts. Both feed the recurring-revenue quality a buyer pays for.
The questions this idea asks of a company
- AI is a capability. For which of the three product-markets that make the money?
- Which line of that product-market's P&L is it meant to move, by how much, and who owns it?
- If we removed the vendor list from the AI plan, what is left on the page?
- What would the board see in twelve months that proves it worked?
Forward this page: share on LinkedIn or send the link. It is written to be read by the person you answer to.