Almanac

Consultant-focused KB for Microsoft Dynamics 365 Contact Center: implementation notes, gotchas, and configuration decisions beyond the official docs — across voice and digital channels, routing, agent and supervisor experience, Copilot & AI, workforce engagement, analytics, administration and security.

feature-copilot-analytics.mdv1 · history
CurrentApplies to Standalone + embeddedUpdated 2 months agoSource Microsoft Learn

What it does

Dashboard tracking how reps use Copilot features, which features are being used, how often, and across which interaction types (cases, emails, conversations). Primary tool for measuring Copilot adoption post-deployment.

Key facts

  • Not independently enabled: requires Copilot help pane and Copilot summaries both opted in first
  • Up to 24-hour provisioning delay after enabling before data appears
  • Interaction tracking scope is configurable: cases, emails, conversations can be included/excluded
  • Tracks usage, not quality: high usage does not confirm good Copilot output quality

When to use / skip

Turn this on when you enable Copilot. It's the main source of Copilot ROI data. No reason to skip if Copilot is in scope.

Configuration decisions

  • Which interaction types to track: enable all three (cases, emails, conversations) unless there is a specific exclusion reason; granularity is always more useful than aggregate data

Gotchas

  • Enable it with Copilot, not after. 24-hour delay means enabling post-go-live loses day one. Enable during UAT.
  • Usage doesn't equal quality. Pair Copilot analytics with transcript recording and knowledge analytics for the full picture.

Consultant notes

  • Enable Copilot analytics at the same time as the Copilot features themselves, during UAT. Clients who wait until post-go-live to enable it lose the early adoption baseline: and the early weeks are exactly when the "is anyone actually using it?" question comes from sponsors and steering groups. Having data from day one avoids an awkward six-week gap in the adoption story.
  • The usage-vs-quality distinction matters for how you present this in executive reporting. High Copilot usage is a positive indicator of adoption, but it doesn't confirm the outputs were good. If the client needs to demonstrate ROI rather than just activity, pair Copilot analytics with knowledge analytics (to see if Copilot is surfacing relevant articles) and conversation transcripts (to see if suggested responses are being used as-is or heavily edited). That combination gives a more defensible picture than usage alone.
  • Track all three interaction types (cases, emails, conversations) from the start. Clients sometimes scope to conversations only on the basis that "we'll add email later": and then never do, because re-enabling isn't a priority once the project closes. Default to all three unless there is a specific compliance or scoping reason to exclude one.

Source last updated: 2025-05-06 | Check this if: New Copilot interaction types become trackable, or analytics expand

Was this accurate?