Why use conversations
Dashboards show you aggregate trends — Conversations lets you drill into the individual calls behind those numbers. When a containment rate drops or a safety flag spikes, Conversations is where you investigate what actually happened, turn by turn. Use Conversations when you need to:- Audit call quality — Review transcripts and agent decisions to ensure your agent meets quality standards.
- Debug unexpected behavior — Inspect which flows fired, what functions were called, and what the LLM was asked on any given turn.
- Identify knowledge gaps — Spot topics the agent couldn’t handle and annotate them so you can improve the Knowledge area.
- Monitor live calls — Watch active calls in real time during launches or high-traffic periods.
- Support compliance workflows — Export call data for reporting or internal review.
For quick QA workflows using Conversations, see QA and analytics. For performance analysis, see Performance monitoring.
Features
Conversation review
Browse, filter, and inspect individual call transcripts — both live and completed. Export call data for reporting.
Annotations
Flag transcription errors or missing topics during review. Feeds a continuous improvement loop for ASR and your Knowledge area.
Conversation diagnosis
Toggle debugging layers to see conversation variables, flow paths, function calls, topic citations, and turn latency.