Skip to content

Skill Usage Insights

The Skill Usage Insights page summarises how the tenant’s contacts are using Conversation Skills over a chosen time window. It is the tenant-facing analytics view onto skill dispatch — totals, success rate, latency, and a bar chart of the most-invoked skills.

Skill Usage Insights page with date-range presets, KPI cards for dispatches / success rate / latency / error rate, and a bar chart of top skills by dispatch count

  • Route: /communications/insights/skills
  • Menu Path: Communications → Skill Usage
  • Primary audience: Tenant administrators and CX owners reviewing how contacts are using the chatbot.
PermissionDescription
communications.access (Read)View the page
PresetWindow
7dThe last seven days
30dThe last thirty days
90dThe last ninety days
CustomPick a start and end date

The selected preset drives every metric and the chart on the page.

CardDescription
Total dispatchesNumber of skill invocations in the window.
Success ratePercentage of dispatches that returned successfully (no skill error, no budget exceeded).
Average latencyMean time between dispatch and response, in milliseconds.
Error ratePercentage of dispatches that ended in a skill error or other non-success outcome.

A bar chart showing the top ten skills by dispatch count over the selected window. Bars are labelled with the skill’s display label (resolved from the registry’s menu_button labels — for example, finance.ar_balance shows as “AR balance”). Hovering a bar reveals the exact dispatch count.

  • Data is scoped to the active tenant. Cross-tenant analytics are not exposed on this page.
  • The chart resolves skill labels from the skill registry. If a skill has been retired, the chart still shows historic dispatches under the retired skill’s fully-qualified name.
  • Date range queries are limited to the retention window the platform configures for dispatch records (currently 90 days); requests beyond that window are silently capped.
  1. Open Communications → Skill Usage.
  2. Select the 30d preset.
  3. Read the success rate card to gauge overall chatbot health.
  4. Read the bar chart to see which skills are most valuable; consider promoting under-used ones in the Conversation Menus.