Overview
After a simulation call completes, SuperBryn analyzes the transcript and produces a detailed evaluation. Results include quantitative metrics, qualitative LLM evaluation, and the full call recording.Viewing results
- Go to the Jobs page
- Click on a completed job to open the Job Analysis page
What’s included
Call recording
The full audio recording of the simulation call. Listen back to hear exactly what happened.Transcript
Turn-by-turn transcript of the conversation. Each turn shows:- Speaker: AI caller or agent
- Text: What was said
- Timestamp: When it occurred
Metrics
Quantitative measurements extracted from the call:| Metric | Description |
|---|---|
| Latency | Response time between caller input and agent response |
| Interruptions | Number of times speakers talked over each other |
| Silence duration | Total silent time during the call |
| Words per minute | Speaking rate for each participant |
| Speaking ratio | Percentage of time each participant spoke |
| Call duration | Total length of the call |
Step-by-step evaluation
An LLM evaluates the agent’s behavior against the scenario’s expected outcome. Each step in the expected outcome is checked:- Pass: The agent correctly performed this step
- Fail: The agent missed or incorrectly performed this step
- Partial: The agent partially met the expectation