Sample Team Composition Report
Six of twenty-four employees assessed. This is what your employer console produces once at least three people finish.
Composition
Coverage gaps
Nobody can explain AI-assisted work to clients, patients, or executives. The first hard external question will stall the whole programme.
Wins will stay personal. Nothing gets templated, so every improvement dies when the person who made it moves on.
Nobody will catch AI errors. This is the single highest-risk gap on this list — it is a hiring or training gap, not a stylistic preference.
Chemistry pairs
The AIRE Architect can move a whole group; the Analyst makes sure the thing being scaled actually works. Pair them before any team-wide rollout.
The AIRE Pragmatist has the fluency and no standing; the Architect has the standing and can give the fluency a legitimate path.
Friction pairs
Fluency against ground truth. The AIRE Pragmatist moves faster than the record allows; the Steward reads that as carelessness. Give them a shared, written standard.
Throughput against precision. The AIRE Operator will ship before the Analyst is comfortable. Agree the verification budget in hours up front.
Manager talking points
- I assessed all of us, myself included — this is about what each of us is best at with AI, not a performance review.
- Nobody here is 'behind'. The four dimensions measure different working styles; a low score in one is a specific, fixable habit, not a verdict.
- Here is our team shape and the gap it leaves — that gap is mine to fill, either by training or by hiring.
- Nothing AI-produced goes out externally without a named person checking it. Here is who that is for each type of work.
- Everyone gets one measurable thing to try in the next 30 days, and we review results — not effort — at the end of it.
4-week lunch-and-learn curriculum
- — 10 min: live demo — draft one request for information (RFI) and one daily report from voice notes, on screen, unedited then edited.
- — 10 min: the three failure modes on a jobsite — fabricated spec citations, stale drawing revisions, confident wrong quantities.
- — 10 min: each person names one task in their week that takes over an hour and is mostly typing.
- — 10 min: hands-on — everyone dictates a daily report on their own phone and cleans it up.
- — 10 min: what must never be AI-generated without review — anything going to the owner, architect, or inspector.
- — 10 min: set the standard format the whole team will use for the next two weeks.
- — 10 min: revision compare demo on a live job's last two sets.
- — 10 min: AI takeoff vs. manual takeoff on the same area — compare the numbers in the room.
- — 10 min: agree the tolerance and the check before anything is used on a live bid.
- — 10 min: each person reports their measured result — minutes saved, errors found.
- — 10 min: decide which two practices become standard on every job.
- — 10 min: name owners and the review date; write it into the project procedure.