End-user feedback is information about system quality or satisfaction provided by people who interact with a deployed system. In agent evaluation, it is one evaluator category alongside code-based evaluators, LLM judges, and human evaluators. section 10.3
Use end-user feedback as part of a broader evaluation approach:
- Combine it with other evaluation methods when validating agents whose prompts, models, or data change. section 10.3
- Treat it as a signal grounded in people’s interactions with the deployed system. section 10.3
