Results
Every figure from a run. Measured, not promised.
Each workflow below was designed through the product from an expert's description, then scored against a known answer on data its tuning never saw.
Tested workflows
Measured, not promised.
Bank reconciliation
0settled wrong
108 of 108exceptions surfaced
176 → 7judgement calls sent to a person
UK VAT return for a small business (pub VAT)
0settled wrong
15 of 15problems reached the preparer
67 → 25items for people after the run
Custody reconciliation
0settled wrong
77% → 97%lines typed right
47% → 58%lines settled without a person
Everything we've run
Settled wrong: zero.
| Test | Settled wrong |
|---|---|
| Bank reconciliation, nine held-out monthsWording tuning never saw · 108 real exceptions, all surfaced | 0 |
| The same months, three different judgement modelsThe Judgement Engine and two frontier models in the same seat | 0 |
| Against a 1,154-line program an AI wrote for the jobThree unseen months · 36 of 36 exceptions | 0 |
| Custody statements from a live fund42 real lines first seen after tuning | 0 |
| UK pub VAT returns, a pub held out of tuning105 judgements · 15 of 15 problems reached the preparer | 0 |
| Deliberately wrong picks at 100% confidenceFed to a changed check before it went live · 36 of 36 rejected | 0 |
| Total | 0 |