Source Files
| File | Type | Rows | Period | Refresh Date | Used For |
|---|---|---|---|---|---|
| CopilotActivityUserDetail (90d) | M365 Admin — Copilot Activity | 88 | 90 days to 2026-03-31 | 2026-03-31 | Division Report (prior period data) |
| CopilotActivityUserDetailV46_12_2026_6_42_15_AM.csv | M365 Admin — Copilot Activity | 93 | 180 days to 2026-06-07 | 2026-06-07 | Primary source — all current reports |
Fields Used (180-Day Primary Export — June 2026)
| Field Name | Used For | Notes |
|---|---|---|
| Report Refresh Date | Period identification | 2026-06-07 for current primary export |
| Report Period | Period length (days) | 180 for primary export |
| User Principal Name | Unique user identifier | 5 x_-prefix users are likely guests/service accounts; included in counts but flagged |
| Display Name | User display in tables | |
| Last Activity Date | Most recent active signal · sub-period trend analysis | Blank for users with 0 prompts in period |
| Prompts submitted for All Apps | Primary usage metric · segmentation · ranking | Sum of all Copilot prompt interactions in period |
| Prompts submitted for Copilot Chat (work) | Chat work breakdown (top users only) | |
| Prompts submitted for Copilot Chat (web) | Chat web breakdown (top users only) | |
| Active Usage Days for All Apps | Habit frequency metric | Count of days with any Copilot activity in period |
| Last activity date of [App] (UTC) | App adoption signal · sub-period trend | 12 app columns; non-null = used in period at some point |
| Division | Division segmentation | Not present in 180-day export. Division Report uses 90-day prior export only. |
Metrics Calculated
| Metric | Definition |
|---|---|
| Active users | Users with Prompts submitted for All Apps > 0 |
| Not using | Users with Prompts = 0 in the reporting period |
| Regular users | Users with 1–499 prompts (fixed threshold for cross-period comparability) |
| Power users | Users with 500+ prompts (fixed threshold) |
| Median prompts | 50th percentile of prompt distribution among active users only |
| P90 / power-user threshold | 90th percentile of prompt distribution among active users; top 10% of active users by volume |
| App adoption rate | % of licensed users (93) with a non-null last activity date for that app. Cumulative — not limited to current window only. |
| Weekly prompt rate | Total prompts ÷ (period days / 7) — used for cross-period normalised comparison |
| Avg active days | Mean of Active Usage Days for All Apps across all active users |
| Sub-period activity | Users whose Last Activity Date falls within each 60-day band of the 180-day window. Represents last recorded activity date, not activity volume within the sub-period. |
Trend Methodology
Two 180-day snapshots are now available (ending 10 May 2026 and 7 June 2026). Direct metric comparison is valid with the caveat that the windows overlap by 152 days — only the 28-day margin represents genuinely new data. Sub-period trends within the current window are built by splitting the 180-day period into three 60-day bands and counting users whose last activity date falls within each band. This is a cumulative adoption signal, not a prompt-volume trend.
Known Limitations
Division field missing in 180-day export. The current primary export does not include a Division column. The Division Report uses the prior 90-day export (ending 31 March 2026). To refresh division analytics, export a Copilot Activity report that includes the Division field from M365 Admin Center.
Prompt volume by sub-period not derivable. The 180-day export provides aggregate totals only. Monthly or sub-period prompt volume trends require a separate 30-day export per period. This is the most significant analytical gap in the current data model.
App adoption is cumulative. Last-activity dates indicate usage at some point in the 180-day window, not necessarily in recent weeks. Figures should be read as "has used this app during the period", not "actively using now".
Rolling 180-day window overlap. The two 180-day snapshots share 152 days of data. Metric changes between snapshots reflect the effect of removing the oldest 28 days and adding the newest 28 days — not a month-over-month measurement.
x_ prefix accounts. Five users have x_-prefixed or onmicrosoft.com email addresses. These may be service accounts, contractors, or transitional identities. One (Michael Garrett) is the sole non-user. Verify before including in enablement outreach or licence decisions.
App-level prompt counts partially available. Only Copilot Chat (work) and Copilot Chat (web) include explicit prompt counts in the export. All other app surfaces are represented by last-activity date only — individual app prompt volumes are not available.
Refresh Instructions
How to refresh this report package
1. Export a new CopilotActivityUserDetail report from Microsoft 365 Admin Center (Reports → Usage → Copilot). Use the 180-day period option for the standard refresh.
2. For prompt-volume trending, also export a 30-day version and upload both files. This unlocks monthly prompt rate trends that cannot be derived from the 180-day aggregate alone.
3. Upload the CSV files. Reports will be regenerated using the same design, navigation, filenames, and metric definitions.
4. To restore Division data, export a version of the Copilot Activity report that includes the Division column — available in some M365 export configurations. Upload alongside the 180-day file.
5. A PowerPoint executive report can be created on request after any refresh.