Here is what we found.

A shortened example: a fictional bank’s customer chat assistant, set out the way every report is.

How our report is structured.

  1. Executive summary

    What we tested, the findings at a glance, and what to change first. One page your leadership can read.

  2. Scope and approach

    What was in scope and what was not, the rules we tested under, and how severity is scored from likelihood and impact.

  3. Findings

    One page each, by severity: what happened, how to reproduce it, the conversation itself, and its remediation.

  4. Appendices

    The severity rubric, the retest, the articles of the EU AI Act and the General Data Protection Regulation (GDPR) a finding may bear on, and the measurements behind every rate.