Insights & Analytics
Insights gives you visibility into how users interact with your agents — what they're asking, what they're trying to accomplish, and where conversations are going well or falling short.
Accessing Insights
Click Insights in the left nav. View data across all agents or filter to a specific one. Use period filters (24h, 7d, 30d, 90d) to narrow the timeframe.
Conversation Metrics
Overview
- Total conversations — with trend comparison to previous period
- Completion rate — percentage of conversations that reached a terminal node
- Average messages — messages per conversation
- Leads captured — total lead capture events
Conversations Over Time
Daily conversation counts plotted on a chart for spotting trends and spikes.
Top Agents
Agents ranked by conversation volume — see which agents are getting the most use.
Goal Performance
Aggregated goal completion by category — see how well agents are achieving their objectives.
Funnel
Visualizes the conversion pipeline:
- Started — conversation initiated
- Engaged — 2+ messages exchanged
- Progressed — 3+ journey nodes visited
- Converted — lead captured
- Completed — conversation finished successfully
Flow Dropoff
Per-node visit and dropoff rates — identify exactly where users leave the journey.
Knowledge Gaps
Questions your agent couldn't answer, normalized and grouped. Use these to improve your knowledge base.
Reusable Answers
Frequently-repeated long assistant answers — candidates for adding to your knowledge base as Q&A pairs.
User Insights
- Returning visitors vs new
- Average session duration
- Peak usage hours
- Traffic sources
Journey Effectiveness
Flow completion rates and bottleneck detection — which nodes slow users down.
Document Metrics
For agents that generate documents: total generated, average sections per document, average revisions.
Automation Metrics (Action Agents)
When you have Action agents, an additional Automation section appears:
Run Overview
- Total runs — with period comparison
- Success rate — percentage of runs that completed without error
- Average duration — how long runs take
- Credits used — total AI credits consumed
Runs Over Time
Daily run counts broken down by status (completed vs failed).
Top Action Agents
Agents ranked by execution volume with success rate and average duration.
Recent Failures
Failed runs with error details — use these to debug and fix workflow issues.
Conversation Review
The Conversations page in the left nav provides a review queue system for managing conversations that need attention:
- Abandoned — user dropped off mid-conversation
- Negative — negative sentiment detected
- Intent — specific intents tagged for review
- Lead — lead information captured
- Handover — user requested human handoff
- Assigned to me — handovers assigned to you
- Unassigned — handovers waiting for assignment
- Knowledge gap — agent couldn't answer the question
- Reusable answer — long, high-quality answer worth saving
Each conversation has a review state where you can flag issues, assign to team members, and add internal notes.