What your teams found in their own work, what they built with it, and what it's worth going forward — not a completion record.
reclaimed every month, self-reported by the people doing the work — up from a baseline of essentially zero five weeks ago.
People had to name the work that was actually wasting their time, and then a meaningful share of them had to go build something about it — not just attend.
distinct manual tasks identified across 7 functions, in participants' own words, before any AI instruction began.
of those tasks already have a working first version, built by a participant during the cohort — not a hypothetical use case.
Measured the same way at intake and at close: a scored knowledge test, a graded prompting exercise, and self-reported usage — weighted so self-report alone can't move the number.
Ranked by reported hours per person per week. Full inventory of 31 tasks available on request — this shows the highest-impact quarter of it.
Cross-checks invoice lines against the PO and flags only mismatches.
Reformats the base technical spec per customer profile in one pass.
Turns raw handover notes into the daily plant report automatically.
We don't report who attended or who's "behind" — that reads as a performance record of your own staff. What we track: how many people went past the required work into something optional.
Ranked by hours saved and by how many people named the same problem independently.
None of these three needs a developer to start. This is a validated starting list, not a finished business case.
Task inventory and time-saved figures are self-reported and aggregated — directional, not accounting-grade. The readiness score blends a structured knowledge test and a graded applied exercise with self-reported usage, weighted so self-report can't move the score alone. No individual's score, completion status, or attendance appears anywhere in this report, or in anything shown to a participant's manager — this cohort's sponsor is also several participants' direct manager, and a completion metric under those conditions reads as staff evaluation, not a training outcome.