How Ask Greg's quality is evidenced | ATX AnalyticsTable Docs

How Ask Greg's quality is evidencedWhat the live capability matrix tests, the latest result, how numbers are re-checked against the database, what Ask Greg refuses on purpose, and how to run the matrix against your own model. Unit tests of an AI assistant usually script the model. The live capability matrix does not: it sends 44 real requests to a real model endpoint and judges what comes back by behaviour — which action Ask Greg proposed, whether it validated, and, for every proposal that contains SQL or an aggregate, whether running it on a seeded database gives the number the fixture was built to give.

What this page covers

Latest result
What the 44 cases cover
Guarantees that hold whatever the model says
Run the matrix against your own model

Settings and controls documented

Azure OpenAI adapter
OpenAI-compatible adapter
Answering with SQL
Charts
Pivot, KPI, dashboard
SQL, cells, value lists, whole project
Transforms and pipelines
Reports and schedules
Alerts and e-mail
Insights
Refusals

Full page

The full page — procedures and the explanation of every setting — is available to signed-in users.

Glossary terms on this page

KPI dashboard: definition

ATX Institute on LinkedIn

This site needs JavaScript for the full experience. Every page above is reachable without it.