How to Evaluate LLM Timeout Events

1. Open call timeline.

2. Identify:

   • Delay between caller input and system response.

   • Retry attempts.

   • Timeout trigger.

3. Confirm transfer reason.

4. Compare frequency across similar calls.

Multiple timeouts in short window may indicate provider issue.

Retry logic:

• Attempts recovery

• Is capped to prevent looping

• Escalates after threshold exceeded

1. Filter calls by agent_failure.

2. Open timeline.

3. Confirm:

   • Retry attempts occurred.

   • Retry limit enforced.

   • Transfer occurred after limit.

Endless retry should never occur.

If transfer seems premature:

1. Review transcript.

2. Check for safety language.

3. Check duration limit.

4. Check language instability.

5. Confirm routing configuration.

6. Compare against other calls.