For CFOs and CEOs

Should we stop this AI pilot or keep going?

Decide whether to stop an AI pilot on three questions: what the pilot has cost so far, what it would take to finish, and whether anyone will sign for the return. If the third answer is no, stop. More time will not produce an owner.

A pilot rarely fails loudly. It runs past its date, the team is busy, the demos look fine, and the review keeps moving to next quarter. I keep seeing pilots in that state for a year. By then it is hard to stop, because stopping looks like admitting something.

It does not have to. A pilot that taught you the idea does not pay has done its job. When was yours last reviewed against a number?

Run a short check on this project

A two-minute video and a check you can fill in with rough numbers.

Several pilots and no clear picture?

The working session on what AI spend returns and what to stop paying for.

What to look at first

01

What has it cost, with people

Licences and usage bills, plus the hours of everyone who worked on it. Money already spent should not decide the next step, but you need the figure to see the monthly cost of waiting.

02

What it takes to finish

Count what it takes to run in the business every day, with someone responsible for it. If that number is larger than the return, the decision is made.

03

Who signs for the return

One person whose budget shrinks or whose revenue grows if it works. With a name, keep going and set a date. Without one, stop or change it.

The longer version: We announced an AI budget. What return should we hold it to?