AI Dev Check-in
A fast read on what a development organization is using AI for today, and the starting point for deciding what to change.
Very few engineering organizations can describe their own AI practice with any precision, and you cannot improve a practice you cannot describe.
Background
Engineering organizations adopted AI tooling faster than they built any shared view of it. Ask five engineers in the same team what they use and where it helps, and you get five different answers, all of them true.
That makes improvement hard to plan. Without a baseline, a decision about tooling, training, or standards is a decision about anecdotes. This gives a team the baseline: what's in use, where it earns its place, and where it doesn't.
It's deliberately a check-in rather than a scorecard. Nothing here produces a grade, it is a baseline - a starting point for improvements.
How it works
Two tracks, so length isn't an excuse to skip it. Express is ten questions in three or four minutes, for a team that wants a picture this afternoon. Comprehensive runs about thirty-five questions in seven to thirteen minutes, for a group ready to look properly.
Both cover the same ground: which tools people reach for, what they use them on, what's working, and what got dropped after the first month. The output is a starting point for a conversation, and a result showing very little adoption is as useful as one showing a lot.
Technology
A hosted web app with a terminal-style interface, deliberately plain, so it reads like an instrument, not a marketing page. The build stack is being revisited as it moves off its first working version.
How it was built
Spec-first: the question set came before the build, since the questions are the product and the form around them is not. What exists today is a first working version, and the production build hasn't happened yet.