Prep the pilot run with evidence-capture instrumentation
Packaged know-how that tells an agent how to do a job well.
You might say…
“I ran three pilots before I started instrumenting them — by the time I'd noticed a pattern, I had no notes to point to.”
What it does
Produces a pilot run-sheet mapping each method step to what evidence to capture and which failure signals to watch for, so the pilot generates structured evidence about the method rather than just serving the case.
Trigger: Use before the pilot delivery starts, to structure what you observe and capture during the run.
I/O: Method step sequence + templates + rubrics → pilot run-sheet with per-step evidence-capture instructions and failure-signal watchlist
Recognise the problem?
The primitives are the commodity part. The fastest next step is a conversation about composing them into something that works for you.
Start a conversation