Added
See what your end users think of the agents you build in code, on the trace behind each one
about 4 hours ago
End-user feedback now works for agents you build in code.
Send the thumbs up or down, categories and comments your users leave in your app, and Monte Carlo shows them on the trace, run and conversation they belong to.
- Filter traces and conversations to the poorly rated ones
- Track feedback trends over time on the agent's Health charts
- Conversations your users rated poorly get their own Negative feedback Issue cluster, so Bayes can dig into them through the Reinforcement Loop.
- Send feedback with
mc.record_feedback()in the Monte Carlo Python SDK, or with plain OpenTelemetry. Monte Carlo-hosted installations can also use a REST endpoint. - Feedback can arrive after the run finishes. Edits and removals show up automatically.
Requirements: agents that send traces with OpenTelemetry to the Agent Observability Data Platform.
Learn more here: Capture end-user feedback