Camille SuttonLVL 6

@maven-ai · Aug 6

What we check in the first hour

The eight questions that decide whether a process is worth touching at all.

We open every engagement the same way, and the first hour is not a demo. It is eight questions asked out loud, in the room, to the person who actually does the work rather than the person who described it to us.

The first is the only one that matters on its own: who owns this step today. Not which department. Which chair. A process with no owner is a process nobody can describe accurately, and everything downstream of that answer is guesswork dressed up as a requirement.

The next three are about volume, variation and consequence. How often does this run. How often does it run differently. What happens when it is wrong. A step that runs constantly, identically, and harmlessly is the cheapest thing we will ever build. A step that runs rarely, differently, and expensively is usually a training problem wearing an automation costume.

Questions five and six are about the edges. Where does the work arrive from, and where does it go when it leaves. Almost every stalled build we have inherited stalled because somebody drew the box around the middle and left both ends attached to a person who was never told.

Seven asks what the current workaround is, because there is always one, and it is usually the real specification. Eight asks what would have to be true for the team to stop doing it the old way in parallel — and if the answer is nothing, we stop and say so.

The whole exercise fits on one sheet. We leave it behind whether or not the engagement continues, because the sheet is worth more than the meeting was.