Accuracy

Measured from reviewer corrections as people work, not sampled quarterly.

Field accuracy

Confirmed over confirmed plus corrected, for fields a person actually looked at. What the model gets right when it was asked.

FieldConfirmedCorrectedAccuracy
Invoice number10100.0%
Issue date40100.0%
Due date10100.0%
Supplier name2166.7%
Supplier tax ID20100.0%
Currency10100.0%
Subtotal10100.0%
Total20100.0%

Auto accept precision

The same ratio over the random sample of fields that were accepted without review. This is the number that answers whether the fields nobody checked were safe to not check.

Precision93.3%
Sample size15
Corrected in sample1
Corrected outside the sample3

3 auto accepted fields turned out to be wrong when someone looked. The threshold is letting errors through.

Time saved

Fields a person never had to touch, times the manual baseline. Only fields nobody looked at count, which is the strictest honest reading.

Time saved19m 0s
Fields never touched76
Manual baseline15s per field

Settings these numbers were judged by

Recorded on each extraction at the time it ran, so tuning them later does not rewrite what already counted as auto accepted.

ThresholdSample rate
0.850.1

Every figure above traces to one query in docs/architecture.md.