Microsoft 365 Copilot costs $30 per user per month. That is $360 per year — added on top of your E3 or E5 license cost, on top of your existing Microsoft agreement. For a 500-person deployment, that is $180,000 annually before you account for the implementation effort, change management, and ongoing enablement required to see meaningful adoption.
The finance question is straightforward: is it working? Most organizations that have deployed Copilot broadly cannot answer that question with data. They have adoption dashboards in the Microsoft Copilot Dashboard (via Viva Insights), and those dashboards show Copilot activity metrics — but activity is not value. A user who runs one Copilot prompt per week is “active” in Microsoft’s reporting. Whether that prompt saved them ten minutes or ten hours is invisible to the platform.
This article is about building the framework to measure Copilot ROI in terms finance teams can act on.
What Microsoft’s usage data tells you — and what it doesn’t
The Microsoft Copilot Dashboard (available through Microsoft 365 Admin Center for tenants with Viva Insights) surfaces these metrics per user:
- Number of Copilot prompts submitted (last 28 days)
- Which applications Copilot was used in (Teams, Outlook, Word, Excel, etc.)
- Whether the user provided a thumbs-up or thumbs-down on responses
- Active Copilot users vs. licensed users ratio
These are activity metrics. They tell you Copilot was touched. They do not tell you whether the touch created value. A user who asks Copilot “summarize this email thread” and gets a response they never act on appears identically in the dashboard to a user who uses Copilot to draft client proposals, automate meeting summaries into action items, and extract data from financial reports in minutes rather than hours.
The second category of user is getting $30/month of value. The first might not be getting $3 worth. Microsoft’s dashboard cannot distinguish between them.
The three-tier Copilot user segmentation
A practical approach to measuring Copilot ROI segments users into three categories based on depth of use, not frequency.
Tier 1: Power adopters. These users have integrated Copilot into workflows that previously took significant manual effort. In finance contexts, this typically means using Copilot in Excel for financial model summarization, Copilot in Teams for meeting intelligence and action item capture, and Copilot in Outlook for managing high-volume correspondence. Usage frequency is high (5+ active days per week) and application breadth is wide (3+ apps). These users are clearly getting value.
Tier 2: Casual users. These users have tried Copilot, use it occasionally for convenience tasks, but have not integrated it into any core workflow. Usage is sporadic (1-2 days per week on average) and concentrated in one or two low-complexity applications. Whether these users are getting $30/month of value is uncertain and depends on how much time they’re saving in those occasional uses.
Tier 3: Non-adopters. These users were licensed and either never activated Copilot or activated it and stopped using it within the first 30 days. They are paying $30/month for a product they are not using. This is the clearest waste signal.
Microsoft’s dashboard will surface Tier 3 clearly (zero or near-zero activity). The distinction between Tier 1 and Tier 2 requires your own framework.
Building a productivity value framework
The challenge with Copilot ROI is that the value it creates — time saved — is inherently difficult to quantify in dollar terms without knowing what that saved time was redirected toward. The most practical approach for finance teams is a sampling methodology rather than instrumentation of every user.
Step 1: Survey Tier 1 users. Ask your highest-usage Copilot users two questions: (a) In an average week, roughly how many minutes does Copilot save you? (b) What do you do with that time? The first question gives a productivity proxy. The second tells you whether the time saving is real (redirected to higher-value work) or phantom (spent on the same low-value tasks more efficiently).
Step 2: Convert to dollar equivalents. Take the average loaded hourly rate for the role category (finance analyst, account manager, executive). If a finance analyst at $95K loaded ($48/hour) saves 45 minutes per day using Copilot, that is $180/week in productivity, or $720/month — 24× the license cost. That ROI case is compelling. If the same analyst saves 5 minutes per day, the math is $20/month — still marginally positive, but not a case for broad deployment.
Step 3: Identify the non-adopters and run an exit survey. Understanding why Tier 3 users stopped matters. Common reasons: the outputs were low quality for their specific use case, they couldn’t figure out effective prompting, their workflow didn’t have a natural Copilot touchpoint. The reasons inform whether the right intervention is training, workflow redesign, or license reclamation.
The license reclamation decision
For users who have been inactive in Copilot for 60+ days, the default recommendation is license removal. The license can be reassigned to someone with higher potential for adoption.
Before removing, however, it is worth checking whether the user went through any onboarding. Copilot adoption follows a steep learning curve. Users who received prompting guidance and use case examples in their first two weeks adopt at materially higher rates than users who were given access with no enablement. A user who failed to adopt after proper onboarding is a reclamation candidate. A user who was provisioned and never told what to do with it is not yet a reclamation candidate — they are an enablement candidate.
This distinction matters because the cost of good Copilot onboarding (a few hours of structured guidance) is trivial compared to the cost of cycling through license assignments and re-onboarding new users repeatedly.
The CFO view
At the portfolio level, a Copilot investment across 500 users at $180,000 per year should be justified by one of two things: a documented productivity return that exceeds the cost, or a strategic capability that cannot easily be priced (competitive differentiation, talent retention, future-proofing). Both are legitimate justifications. Neither is the same as “we deployed it because Microsoft included it in our renewal conversation.”
The measurement framework above is not complex. Running it quarterly and sharing results with the executive team converts Copilot from a line item on the IT budget into a tracked investment with a measurable return — which is the posture every CFO should want from a $180,000 spend.