What we tell CEOs who ask about AI strategy
There is no AI strategy. There is a business strategy, and better tools to execute it with.
- Topic
- AI Enablement
- Kind
- From the archive
- Reading time
- 3 minutes
- Written by
- Telemyr
01 / 06
The question comes up in most first conversations. What should our AI strategy be?
Our answer tends to disappoint for the first five minutes. We do not think a separate AI strategy is a useful thing to have.
02 / 06
Strategy did not change
Your strategy is still about margin, growth, cash, and the customers you serve. AI did not add a new item to that list.
What changed is the cost of executing. Analysis that took a team now takes an afternoon. Software that took a year now takes weeks. Work that justified a hire can be built instead.
A strategy document about the technology itself usually means the real question has been avoided. The real question is which commercial outcomes you want, and whether the new costs of execution change what is possible. Often they do, and that is worth a strategy conversation. The technology is not.
03 / 06
Start from two or three numbers
When a CEO asks us where to start, we ask which two or three numbers they most want to move this year.
Inventory down without losing sales. Margin by product visible and improving. The close finished in three days. Headcount flat through the next phase of growth.
Those are targets a team can organise around, and every one of them implies the AI work underneath it. Pick the numbers first and the technology roadmap writes itself. Pick the technology first and you get pilots.
04 / 06
Then the honest question about foundations
Most companies want to start with the exciting layer. Agents, forecasting, automation.
The honest question is whether your numbers agree with each other today. If two reports on the same subject give two answers, that has to be fixed first, because everything clever inherits the problem underneath it.
The foundation takes weeks now, not the year it used to. But skipping it is the most common expensive mistake we see, and it is usually made because it was nobody's favourite part of the plan.
05 / 06
What the CEO personally has to do
Not much, and the little is decisive.
Name the outcomes and put your weight behind them. Give the work to the people who own the process, with IT supporting rather than owning. Insist the old process is switched off when the new one lands, because that is the decision everyone below you will avoid.
And use the tools yourself, visibly. A leadership team that asks the warehouse its own questions changes the culture faster than any programme.
So our answer to the strategy question is short. Decide what you want the business to achieve, check whether the new cost of execution changes that ambition, and then raise it.
06 / 06