Twelve months into a Copilot rollout, the most common conversation Frontrow has with Australian clients is not about features, it is about measurement. The board or the CFO has seen the licence bill and wants to know whether Copilot is actually being used, and whether the usage is generating the outcomes that justified the spend. Until recently, answering that question required assembling data from multiple admin centre reports and stitching it manually. Microsoft's Copilot Analytics dashboard, now generally available in the Microsoft 365 admin centre, is the answer, with some important caveats about what it does and does not measure.
What the dashboard shows
Copilot Analytics surfaces three primary data categories. Active usage shows the percentage of licensed seats with at least one Copilot interaction in a rolling 28-day window, broken down by app (Word, Outlook, Teams, Excel, PowerPoint, Copilot Chat). Adoption trend shows the week-over-week trajectory of active usage, which is the metric Frontrow finds most predictive of whether a rollout is healthy or stalling. Feature engagement breaks down which Copilot capabilities users are actually invoking, as opposed to which ones the licence includes.
The dashboard also surfaces a Copilot impact section that attempts to quantify time saved, drawing on Microsoft's proprietary Copilot Value Model, a framework developed from telemetry and survey data across Microsoft's enterprise customer base. For each active user, it estimates hours saved per week based on the type and volume of interactions. The methodology is documented in the admin centre's help section, and it is worth reading before citing the numbers in a board report.
What it measures well
Active usage at the seat and app level is reliable and well-granulated. The trend data is accurate. If the dashboard shows Copilot Chat usage climbing week-on-week while Word and PowerPoint usage is flat, that tells an Australian IT team something specific and actionable: Copilot is being used as a search tool rather than an in-document writing tool, and the training program may need to shift emphasis.
The feature engagement breakdown is particularly useful during a phased rollout. When Wave 2 features like Copilot Pages and the Researcher agent appear in the dashboard, usually around two to three weeks after an admin enables them, the engagement data provides early signal on whether those features need additional champion promotion or whether they are landing on their own.
What it does not measure, and why that matters
The estimated hours-saved figure is a model output, not a measured outcome. It assumes an average time-per-interaction based on Microsoft's aggregate telemetry and then applies that to your tenant's interaction count. It cannot measure whether the output the user received was actually used, whether the draft email was sent or discarded, or whether the summarised document led to a better decision. For roles where Copilot value is primarily cognitive, better preparation, clearer thinking, faster synthesis, the model undercounts. For roles where value is primarily operational, fewer keystrokes, shorter drafts, faster reformatting, it may overcount.
The practical implication for Australian businesses preparing a Copilot review for finance leadership is to present the Copilot Analytics dashboard data alongside a structured qualitative sample. Pick 10 to 20 high-usage users, ask them to document three concrete examples where Copilot changed a work output in the last month, and triangulate those examples against the dashboard numbers. That combination is more credible than dashboard figures alone and more structured than pure anecdote.
How to access and configure the dashboard
Copilot Analytics is accessible in the Microsoft 365 admin centre under Reports > Copilot for Microsoft 365. The default view covers the last 28 days at the tenant level. Filters allow drill-down by department, licence type and date range. The dashboard requires Global Admin or Reports Reader role access; the latter is the correct permission level to share with a management reporting function.
One configuration step most AU admins overlook: user-level pseudonymisation is on by default, which means the department-level breakdown is available but named user data is masked. For organisations that want to understand adoption at the individual level, for example, to identify champions or to manage the licence reclaim cycle, a Global Admin can disable pseudonymisation under Settings > Org settings > Reports. This is a governance decision, not a technical one, and should be recorded.
The 30-day measurement cadence Frontrow recommends
- Week 1: pull the baseline. Document active usage percentage and feature engagement breakdown for the current rollout cohort.
- Week 2: review adoption trend. If weekly growth is below five percentage points per week, engage the business unit champion before assuming the licence is the problem.
- Week 4: first structured qualitative sample. 10 users, three examples each, documented. Compare qualitative themes to the feature engagement data from the dashboard.
- Month 2 onwards: report active usage, estimated hours saved (with methodology note), and qualitative highlights to the exec sponsor monthly. Flag any seat with zero activity for 30 days for the licence reclaim conversation.
Try it
Check your overall M365 utilisation
The M365 Usage Tool sits alongside Copilot Analytics to show which M365 services your organisation pays for but underuses, including the capabilities that improve Copilot outcomes when they are properly deployed.
Step 1 of 4
How big is your organisation?
We'll use this to estimate your total spend and scale the recommendations. Change the seat count if you know it exactly.