2 min lesson
Designing the pilot
Match each case in "Designing the pilot" to the signal and response that fit it.
Step 1 of 2
A pilot is not a free trial with a friendly logo. It's an experiment you run to manufacture the evidence that justifies a thousand-seat rollout.
The contract is signed, the org is curious and you have a window of goodwill. Spend it on one team that will generate a defensible story. The pilot's deliverable is not happy users. It's a readout with a baseline, a number that moved and two or three quotes from engineers the rest of the org respects.
Team selection is the first decision and the one people get wrong. The instinct is to pick the easiest, most enthusiastic team. That gives you a result nobody believes, because the skeptics will say "of course the early adopters loved it."
The team of early-adopter AI fans who already use every tool.
They'll succeed, but the win won't generalize and skeptics will dismiss it.
A representative team on a typical stack with a willing champion.
A normal team's success is the result that travels across the org.
The openly hostile team or one on an exotic stack Cursor handles poorly.
A failure here poisons the whole rollout before it starts.
Pick a team with a motivated champion on a mainstream stack, doing the kind of work most of the org does. You want a result that a VP can point at and say "that's us."
Learn more
Full explanation
The Pilot as an Experiment
Interactive diagram. Step through it with the Next and Previous controls below, or Tab to a region to read its detail.
Each gate must clear before the pilot earns the right to become an expansion case.