3.2 Verification ceiling
Completion requirements
Estimate how much AI output your team can responsibly verify per week. Not how much arrives — how much can be genuinely checked against a standard.
Show your working. Hours available, minutes per item, what counts as a real check.
Then compare it to how much you currently generate. If generation exceeds the ceiling, say where the difference is going.
400–600 words · about 45–75 minutes.
How you will be graded — aim for Proficient.
| 1 · Emerging | A guess with no working. |
| 2 · Developing | Working shown, but review time is treated as free or omitted from the comparison. |
| 3 · Proficient — target | A number, with the working shown: hours available, minutes per item, and what counts as a real check rather than a glance. Compared against how much you currently generate, with the difference accounted for. |
| 4 · Exemplary | Also names where the unverified surplus is currently accumulating, and what it looks like when it fails. |