Adoption analytics
Adoption analytics answer the first question leaders ask about AI: who is actually using it, and for what? PromptGlass combines events from every sensor and connector into one view of active users, apps and use cases, broken down by department, location, manager hierarchy and cost center from your identity provider.
Core metrics
Section titled “Core metrics”| Metric | Definition |
|---|---|
| Active users (DAU / WAU / MAU) | Distinct people with at least one AI interaction in the period |
| Penetration | Active users ÷ licensed or eligible employees in the group |
| Depth | Median interactions per active user per week |
| Breadth | Distinct use-case topics per active user |
| Sanctioned share | Interactions in approved apps ÷ all AI interactions |
| Stickiness | WAU ÷ MAU |
- Overview: adoption trend, top apps, sanctioned vs unsanctioned share.
- By team: a heatmap of department × app, with drill-down to sub-teams (pseudonymized by default; see employee privacy).
- Rollout tracker: for a new tool such as Microsoft 365 Copilot, the share of licensed users active in weeks 1, 2, 4, 8 and 12 after assignment.
- Use cases: topics clustered from prompt content (or from metadata when content isn’t stored), such as email drafting, code review, contract summarization, SQL generation and customer reply.
- Champions: teams with high depth and breadth, useful for peer-led enablement. Individual-level views require the
analytics.individualrole.
Minimum group size
Section titled “Minimum group size”To protect individuals, team views hide any group smaller than the k-anonymity threshold (default 5, configurable per region). Smaller groups roll up into their parent.
Exporting
Section titled “Exporting”promptglass export adoption --group-by department,app --period weekly --since 2026-01-01 --format csvAdoption metrics are also available through the REST API and the MCP server.