Overview

Simulation results show what the simulated user and your agent actually said, plus a Pass or Fail outcome and a reason. Open the run to investigate the conversation that produced the result.

When to use it

Check whether the generated conversation met its success criteria.

Use the simulation result to understand whether the generated conversation reached its success criteria. It helps distinguish an agent behavior problem from a scenario that lacks enough context or an expectation that does not match the intended outcome.

How it works

Review the generated exchange, evaluation reason and logs.

The conversation view renders the generated messages and the simulation’s evaluation reason. The same run panel also provides logs. Overall run status and the result of a particular simulation are different levels of information: inspect the individual outcome rather than assuming that opening or finishing a run means it passed.

Example

Compare a failed delivery conversation with its saved expectations.

A delivery scenario may fail because the agent never asks for the information required to find the order. The useful evidence is the actual exchange and its evaluation reason, compared with the saved success criteria. Review the criteria and agent behavior before retrying.

Next steps

Use the up and down arrow keys to select a result, Enter to open it, and Escape to close the search.