#026 - What has to stop for this to start?

Can you have two girlfriends at the same time?

Technically, perhaps.

Ethically, not in my world.

Practically, it usually ends in disaster.

Not because two is a difficult number, just because both relationships compete for the same finite resources: attention, time, care, and energy.

A serious relationship is not only about your feelings. It requires you to care about whether the other person feels safe, seen, and loved. Then you have to turn that care into action: being present, making time, remembering the small things, planning something special, and sometimes buying a gift.

You cannot promise full commitment twice and expect both relationships to thrive.

We understand this in everyday life, so we say: “One door closes so another can open.” We consider that healthy, so why do we behave differently with AI projects?

I regularly see companies declare ten or twelve AI priorities at the same time. Each one receives a sponsor, a workshop, a slide, and just enough resources to remain alive.

But when everything is a priority, nothing is protected.

In my strategy methodology, there is a specific day when the candidate list becomes a portfolio. We may begin with seven or ten serious ideas. I prefer to finish with two or three committed AI bets.

And we discuss openly what will be killed or deliberately postponed so those bets can be born.

This is not theatre. And it is not punishment. Some of the projects we stop are good ideas. They are simply wrong for this moment, weaker than another opportunity, or impossible to support without starving the work that matters more.

For every initiative selected, I ask one question:

What has to stop for this to start?

If nobody can answer, the initiative is not yet a priority. It is an additional wish.

That conversation requires calm and courage - calm makes room for evidence, and courage makes the trade-off explicit.

Calm plus courage brings clarity. Remember.

I turned this question into a one-page AI Project Kill Scorecard.

Score each candidate from 0 to 2 on three questions:

1. Is this a validated €1M problem?

2. Do we have the future-back data or a fast path to it?

3. Is there a credible 90-day path to production through a real workflow?

A total below 4 means kill or deliberately postpone. A total of 4 or more makes an initiative eligible, not entitled, to commit.

Then force one verdict: commit, deliberately postpone, stop pretending, or reclassify. For every commitment, name what stops or moves.

The scorecard can structure the decision, but the decision itself remains a leadership responsibility.

DOWNLOAD THE AI PROJECT KILL SCORECARD 

Your move.