Power BI lenses to measure the value, adoption, maturity, behaviour, and impact of Microsoft 365 Copilot.
Set up your Copilot data once. Each lens then reads the same model to examine value, adoption, readiness, maturity, behaviour, leadership, and Copilot Studio agents.
🚧 Status: early build. CopilotScope reorganises the existing AI Business Value Dashboard into a kit of named Power BI lenses. See Latest release for the milestone plan.
Most lenses come from the same core setup. You only add a data source to unlock a specialised lens.
| Data source (set up once) | Lenses it unlocks |
|---|---|
| Audit log (core) | ValueLens · AdoptionLens · MaturityLens · LeaderLens |
| Org data + manager hierarchy | LeaderLens roll-ups · org slicing across all lenses |
| Licensed users | ReadinessLens |
| Credit / cost (PPAC / MAC) | ValueLens (cost pages) |
| Copilot Studio transcripts (Dataverse) | StudioLens |
| M365 work behaviour (Purview UAL) | BehaviorLens (WIP) |
| Product feedback / Agent 365 | GovernanceLens |
On Scope Home, each lens tile shows as active (its source is loaded) or greyed with "add
<source> to unlock" — so you literally see which lenses your current setup powers. See
Set up once and docs/PREREQUISITES.md.
Every lens is source-gated — it lights up only when its data is present. Tags: (Preview) = built, still stabilising; (WIP) = in progress.
Cost vs value, ROI, and time & tasks saved — the FinOps view of Copilot. Turns audit activity into human-equivalent hours and money, and (with credit/cost loaded) reconciles that against PPAC / MAC spend.
More information on ValueLens.
Post-adoption usage, reach, and activation — who is active, how widely Copilot has spread, and how engagement trends over time. The core "are people actually using it?" lens.
More information on AdoptionLens.
Pre-adoption readiness — who is licensed, enabled, and ready, plus provisioning gaps and reclamation candidates. Helps you close the gap between licences bought and value realised.
More information on ReadinessLens.
The adoption proficiency curve (Initial → Efficient) with a Maturity Meter gauge. Shows where your population sits on the maturity journey and how it is progressing.
More information on MaturityLens.
Team and function roll-ups with gamified leagues, streaks, badges, and leaderboards — powered by the org / manager hierarchy. Gives leaders a competitive, motivating view of adoption across their org. Shows real names by default; a de-identify toggle is available as an opt-in.
More information on LeaderLens.
Employee work-behaviour patterns (Exchange / SharePoint / OneDrive / Teams) mapped to the AI-activity taxonomy — the manual-vs-AI baseline that shows where Copilot could take work off people's plates. Optional source (Purview UAL may not be available in every tenant).
More information on BehaviorLens.
Deep Copilot Studio agent evaluation — quality, topics, transcripts, errors, feedback, and PPAC credits. Available as a page-group inside CopilotScope and standalone (incl. Dataverse-Direct): StudioLens-for-Copilot-Studio.
More information on StudioLens.
Agent health, governance, and feedback — a first-class governance view for admins and CoE teams: agent inventory, adoption/health signals, and sentiment from product feedback.
More information on GovernanceLens.
CopilotScope reuses one data model and one refresh across every lens (see Product architecture). You set up the shared ingestion once, then lenses appear as their sources load.
- Pick a deployment path — Fabric or SharePoint (see Deployment paths).
- Run the core ingestion — the audit-log ingester (Fabric notebooks) or PAX extract (SharePoint).
- Add optional sources to unlock specialised lenses — org/manager hierarchy, licensed users, credit/cost, Copilot Studio transcripts, M365 work behaviour, product feedback.
- Open the template and set the
Enable_*parameters for the sources you loaded.
Prerequisites (licences, Fabric capacity, Graph / Dataverse / Purview permissions, credit/consumption
notes) are documented once in docs/PREREQUISITES.md — referenced from every
setup section rather than repeated inline.
- Fabric setup:
1. Fabric/README.md(seeded in M1b) - SharePoint setup:
2. SharePoint/README.md(seeded in M1b)
Every optional source follows the optional-source pattern (
EmptyTable+try…otherwise+Enable_*toggle), so measures keep binding whether or not a source is present.
| Path | Best for |
|---|---|
| 1 · Fabric | scheduled Spark ingestion, larger volumes, PPAC message-credit pages |
| 2 · SharePoint | flat-file / Power Automate (PAX) landing, no Fabric capacity required |
Both paths ship the same lenses over the same semantic model — only the ingestion differs.
- BehaviorLens depends on Purview UAL / M365 work-behaviour data, which may not be accessible in every tenant. The lens is optional and greys out when the source is absent.
- The manager / org hierarchy that powers LeaderLens roll-ups is currently produced only on the SharePoint (PAX) path. The Fabric org ingester builds it in a later milestone (M2).
- Some lenses are (Preview) / (WIP) — see The lenses for current status.
CopilotScope is in early build. The milestone plan (M0 → M6) lives in the engineering brief.
Tip: to be notified of releases, click Watch → Custom → Releases at the top of the repo.
- StudioLens-for-Copilot-Studio — the standalone Copilot Studio agent deep-dive (also runs Dataverse-Direct, no Fabric required). Cross-linked from Scope Home; the same StudioLens page-group also lives inside CopilotScope.
How CopilotScope differs from the Power CAT Copilot Studio Kit. The Power CAT Kit's Agent Insights Hub and Agent Value Summary are Dataverse-solution analytics for makers. CopilotScope is Power BI value / adoption analytics for leaders and CoE teams across all of Microsoft 365 Copilot — the two are complementary, not competing.
CopilotScope is created and maintained by the Microsoft Business Value Advisory (BVA) team.
MIT — see LICENSE.