RL
RL Studio

Rise Data Labs

DashboardWorldsEnvironmentsTask designerSessions
  1. Environments
  3. Live chat routing
  5. Calibrate bot confidence

Calibrate bot confidence

hardMax 100 pts · 0 sessions
Task prompt
Pull chats where bot confidence <0.4 in last hour, label top failure intents, submit calibration batch.
Success criteria
  • Call environment_runner (55 pts)
  • Within 30 tool calls (45 pts)
Run session

Only tasks in the Email environment start sessions. Open a task there to run.

Previous sessions

No runs yet.