Deciding
Give the board something to click
4 min read · 18 July 2026
Item four on the agenda, the AI initiative. Forty slides, twenty minutes of discussion, three opinions. Afterwards the minutes hold the most expensive sentence a project knows: “To be taken up again at the next meeting.”
That sentence costs more than any quote, because for a month everyone keeps working as before and the question afterwards is still exactly the same. Postponing rarely happens out of laziness. It happens because nothing is on the table for the decision to hold on to.
Why the slide doesn’t decide
A slide describes the initiative, but nobody in the room has used it. So every objection is a guess, and among guesses, the loudest or the most cautious voice wins. Neither has anything to do with the initiative. On a prototype, the nature of the objections changes. “I don’t think anyone needs this” becomes “at this step I would drop out”, and you can talk about the second sentence, because it names a spot.
That’s how we built it for a large Swiss corporation. A decision paper on an AI programme, and every use case in it is not a picture but a small clickable prototype, playable step by step, on a laptop and on a phone. Whoever has to decide can first try out what they’re deciding on.
Where the evidence comes from
Trying it in the boardroom is half the way. The other half comes from people who aren’t in the room. Jakob Nielsen showed in 2000 that five test users already uncover around 85 percent of usability problems; further users mostly find the same problems again. For a decision, that means no study with a thousand participants is needed. Five real users work through the prototype under observation. That’s all it takes for the meeting to hold opinion against observation instead of opinion against opinion.
A postponed decision costs a month. Five users on a prototype cost a few days.
How we work Prototype sprint
You come with the question your executive team has to decide. We build the case as a clickable prototype and put it in front of five real users. You walk into the meeting with something everyone in the room can play through themselves, and with the record of what five people did with it.
Whether the decision is then yes or no is yours. That it gets made is the point of the exercise.
Item four, the same meeting, a different table. On it a prototype everyone used for five minutes beforehand, next to it the notes from five user tests. Maybe the decision is no, and that too is a good outcome. But the sentence “to be taken up again” has a hard time once people could touch the thing. For the minutes there’s readme, our own tool. It turns the recording of the meeting into the finished minutes.