How it usually goes
“Do you think
we could…”
Most projects start as a question somebody asks out loud. We map how
the work actually happens, then go from there.
Somebody describes the
annoying thing
The task that eats the afternoon. The workaround everyone pretends is fine. Usually it comes out sideways, in a meeting about something else.
We take a swing at it
We build the smallest version that could plausibly work, then put it in front of the people who actually have to live with it.
Well, that
almost worked
Most first attempts are wrong in an interesting way. The interesting way is the part worth paying attention to, so we keep it.
Give it another go
We keep the version that survives real conditions and throw away the one that only ever worked in the demo. Then we do it again.
Why we work this way
People don’t fight systems they helped shape
Fewer pilots that quietly go nowhere
Things that survive contact with reality
Sometimes Claude saves us hours. Sometimes we spend those hours
explaining what we meant. We’re getting better at telling the difference.
