Twelve Rounds of Persona Testing. Four Fixed the Wrong Thing.
By DigiEntropy Engineering · 2026-08-09 · 12 min read
We pointed 18 goal-driven persona agents at one new feature and kept going until they could use it. Here is every round, including the four we spent fixing a bug that lived in the measuring instrument.
About this post — Authored by an AI assistant using DigiEntropy's polynomial corpus, the universal compressor predictor, and the project's chart-generation tooling. Charts are produced by Python scripts that read the same database the live site queries; tables and formulas are pulled from the same engineering modules. Findings reflect the project's current dataset and methodology — send corrections or deeper questions to admin@digientropy.com.
We built a feature called Auto System Design: describe a duty, and the platform picks a real compressor, a real condenser, a real evaporator and the valves between them, from the actual catalogs, with a reason attached to each pick.
Then we pointed a room full of imaginary users at it and did not stop until they could use it. That took eleven rounds — and a twelfth after we redesigned the thing.
This is the log. It is not a success story with the awkward parts filed off — four of those rounds were spent fixing a defect that turned out to be in our own measuring instrument, one was a real person finding something all eighteen personas structurally could not, and the last one caught us putting a wall back up during a redesign.
6 → 20 - endpoint calls, round 10 vs 11 - only the observer changed
Endpoint calls went from 6 to 20. Round 11 ran the same persona, the same missions and the same product code, with only the observer fixed.
0.167 - first non-zero contract score - in eleven rounds
And the first non-zero contract score in eleven rounds, 0.167, appeared.
4 rounds - chasing a defect in the observer - rounds 4, 6, 7 and 9
What it cost, stated plainly: four rounds of product changes chasing a defect in the observer.
index 97 - where the capacity input sat - the page summary stopped at 70
100 visible interactive elements with the dialog open; the capacity input sat at index 97; the cap was 70.
0.0 - mission success after the redesign - round 12, eight personas
Every session failed. Mission success 0.0.
43 - backend tests hold the behaviour - verified from an empty canvas in about 40 seconds
Verified from an empty canvas in about 40 seconds: three slots filled from the real catalogs, each with a stated reason, the one brand compromise labelled as a fallback, and 43 backend tests holding the behaviour in place.
How one round works
The Lab is six layers, and only the middle one costs a model call per action.
Walk one round through the steps below:
Persona - who is using the site
…