Analytics

Analytics tells you whether the agent is getting better, not just busier.

What You Can Measure

The Analytics page combines several categories of insight:

• conversation volume and channel mix
• escalation rate and status breakdown
• AI latency and retrieval timing
• voice session health
• knowledge base coverage and document processing
• search success metrics
• user growth and behavior
• language distribution

Dashboard vs Analytics

Use Dashboard for the quick daily view.

Use Analytics when you need to explain why something changed, compare trends over time, or break a problem down by channel, latency, retrieval, voice, or users.

Sections To Pay Attention To

Conversation health

Use conversation totals, channel breakdown, and escalation rate to understand operational load.

Escalation rate matters because it tells you whether AI is resolving more work or pushing more work to humans.

AI performance

Watch total latency, retrieval latency, cache hit rate, token usage, and model breakdown.

If retrieval latency is healthy but total latency is still high, the bottleneck may be generation or tools instead of source search.

Knowledge base performance

Review source counts, document processing status, recent source updates, and search success.

User behavior

Track total users, new versus returning users, peak activity, and language mix.

Good Questions Analytics Can Answer

• which channel creates the most support load?
• are escalations increasing after a recent change?
• is retrieval getting slower as the knowledge base grows?
• do voice sessions succeed often enough to keep voice enabled?
• are users asking in languages we are not configured for well?

Best Practices

• compare analytics with real conversation review, not in isolation
• watch trend changes after updating sources, settings, or interceptors
• investigate latency spikes before scaling traffic
• use analytics to prioritize improvements, then verify them in Conversations

Related Docs

Conversations

Review chats, reply as human, and inspect voice sessions.

Feedbacks

Review low-rated chats and comments to find quality issues.

Usage & Costs

Track provider spend across AI, voice, and ingestion.