Analytics

See Call history, analytics & simulation for the full model. This page is the practical guide to using it day to day.

Reading a call

Open Call History, pick a call, and use View → Explore to see the path the call took through the flow, the transcript grouped by node, a timeline of variable reads/writes, and per-turn latency. If the flow ran a post-call analysis pass, the call detail also shows a summary, topic, sentiment arc, whether the caller’s goal was reached, and any out-of-scope requests — plus survey results if the flow’s survey was enabled.

Reading the Journeys view

Analytics → Journeys aggregates path data across every call into a flow-wide heatmap: thicker edges mean more traffic, node color indicates drop-off. Use the top-paths table and drop-off ranking to find where callers are abandoning the conversation, then jump back to the flow editor to fix the offending node.

Running a simulation

Before shipping a flow change, go to the flow’s Test stage and launch a simulation batch with a preset (Quick, Standard, or Deep — see Simulation). Review the resulting heatmap and judge findings, and use Improve flow with AI on the batch’s detail page to hand those findings straight to the editor’s AI assistant.