Skip to main content

Consumer Feedback

Consumer feedback captures observations from people, teams, systems, or downstream Product owners who consume an AI Product version.

Feedback may include:

  • qualitative feedback;
  • structured satisfaction/usability feedback;
  • complaints;
  • feature requests;
  • unexpected behavior reports;
  • adverse outcome reports;
  • explainability or transparency concerns.

Feedback is evidence, not truth by itself​

Feedback must retain provenance and context. A single complaint, rating, or popularity score must not directly mutate Product Family, APT, risk, or governance state.

However, feedback can become a material assurance signal when it provides evidence of:

  • safety failure;
  • policy breach;
  • systematic bias;
  • failure against declared quality criteria;
  • unanticipated high-impact use;
  • repeated fitness failure.

Such signals should be routed into Continuous Assurance for investigation.

Consumer context​

Feedback should identify the intended-use or consumption context wherever relevant. A Product can be effective for one context and unsuitable for another without contradiction.