Administration
Observability
Monitor conversations, tokens, latency, and errors across every app.
Observability (Solutions sidebar) is the enterprise-wide view of conversational activity.
The headline metrics
At the top: Total Conversations, Total tokens, Latency P50, Latency P99, and Error rate — the health of your AI operation at a glance.
Conversation logs
Below, every conversation across all apps and channels, with: conversation ID, app name, channel (Web Chat or Voice), status, the user's email, and time. Search and sorting narrow things down; pagination handles large volumes.
Click a conversation ID to open its full content and trace — what the user asked, what the agent answered, and how the interaction unfolded. Use it for quality review, debugging odd answers, and auditing.
Practical uses
- Watch error rates after publishing a new app version.
- Compare channel volumes (voice vs web) to see how users actually reach you.
- Track token totals to keep an eye on model spend.