We Bought Microsoft Copilot Licenses and Nobody Uses Them. What Now?

If your Microsoft Copilot licenses are going unused, the answer is rarely more licenses or a different tool. Copilot lives inside Outlook, Teams, Word, and Excel, so adoption comes from training people inside those apps on their own email, meetings, documents, and spreadsheets, then measuring usage by app and following up where it stays flat.

Copilot has a problem most AI tools do not: it is everywhere and nowhere at once. There is no single place people go to 'use AI'. The capability is a button in the ribbon of apps people have used for twenty years, and unless someone shows them what that button does to their actual work, most people never press it more than once. This post covers the Copilot-specific version of the problem. For the platform-neutral version, read We Bought AI Licenses and Nobody Uses Them.

Why are our Copilot licenses going unused?

The usual causes are predictable, and most stalled Copilot rollouts have several at once:

• Training showed Copilot Chat, not the apps. A launch demo of the chat window does not teach anyone how to summarize the thread they are stuck in or build the pivot table they need by Friday.
• The button is easy to ignore. Copilot shows up inside tools people already know how to use without it. Habit wins unless there is a clear reason to change.
• People are unsure what Copilot can see or what they are allowed to ask. Careful employees hold back when nobody has explained the rules.
• IT owned the deployment, and nobody owns adoption. Licenses were assigned, the project closed, and usage drifted.
• There was no baseline. Without knowing who is fluent and who never started, every fix is a guess.

None of these are problems with Copilot itself. They are enablement gaps, and they show up the same way whether you bought 50 seats or 5,000.

How do we find out who is actually using Copilot?

Start with the data Microsoft already gives you. The Microsoft Copilot usage report in the Microsoft 365 admin center (Reports, then Usage, then Microsoft Copilot) shows enabled users, active users, and the active users rate over the last 7, 28, 90, or 180 days. It also breaks adoption down by app and lists each licensed person's last activity date in Teams, Outlook, Word, Excel, PowerPoint, and other apps.

Two details matter when you read it. First, Microsoft counts someone as active only when they take an intentional action, such as submitting a prompt; opening the Copilot pane is not enough. Second, the per-app view is where the story is. A few things to look for:

• Enabled but never active. People with a license and no last activity date at all. These are your first outreach list.
• Chat only. People who use Copilot Chat but never touch Copilot in Outlook, Word, or Excel. They have the habit but not the workflows.
• One strong app, nothing else. Often Teams meeting recaps. A good foothold to build from.
• Whole teams at zero. Usually a relevance or policy problem, not a skill problem. Talk to the team lead before scheduling more training.

If you also license Viva Insights, the report points to the Copilot Dashboard there for readiness, adoption, and impact views. Either way, export the numbers now. They are your baseline.

Where does Copilot adoption actually happen?

Copilot adoption happens inside the apps, on real work, one habit at a time. Good starting workflows by app:

• Outlook. Summarize a long thread before replying, draft a response from the thread, and use email coaching on a message that has to land well.
• Teams. Meeting recaps with action items, catching up on a busy chat or channel, and rewriting a message before it goes out.
• Word. First drafts built from existing files, summaries of long documents, and rewrites for a different audience.
• Excel. Asking questions of a table in plain language, getting help with formulas, and spotting trends before a review meeting.

Each of these maps to something a person already does every week. That is the point. Training that starts from 'here is a feature' fades. Training that starts from 'here is the thread you are dreading' sticks, because it saves time the same afternoon.

What does Microsoft provide to help with Copilot adoption?

More than most customers use. Microsoft's Copilot adoption hub has guidance paths for users, champions, leaders, and IT administrators, a prompt gallery, implementation checklists, and a Copilot Champion Playbook. It is free and it is good. If your team has not been through it, start there.

What the hub cannot do is sit with your finance team and rebuild their month-end commentary, or work out why one region adopted and another did not. Microsoft's material is written for every organization. The gap it leaves is the part that is specific to yours: your roles, your documents, your data rules, and your people's confidence. That is the part an enablement program covers.

What should a 90-day Copilot recovery plan look like?

A recovery plan does not need to be large. It needs to be sequenced:

• Weeks 1 to 2: Baseline. Export the usage report, run a short literacy assessment for everyone with a seat, and interview the leads of your lowest-usage teams.
• Weeks 3 to 6: Role-based training inside the apps. Live, hands-on sessions per function, in Outlook, Teams, Word, and Excel, on each team's own threads and files. Identify champions in every group.
• Weeks 7 to 10: Workflow builds. Rebuild three to five high-frequency workflows with the people who own them, and document them so other teams can copy them.
• Weeks 11 to 13: Re-measure and hand off. Pull the usage report again, re-run the assessment, and give champions a playbook for the next quarter.

Alongside the plan, publish a one-page guide on what data is fine to use with Copilot and what is not. Clear guidance turns hesitant people into active ones faster than any feature tour. We measure the baseline and the follow-up with SINA, our AI literacy platform, so leadership sees fluency by role alongside Microsoft's usage numbers.

This is the structure of our Microsoft Copilot training, and it follows the Assess, Educate, Enable, Scale model described on our How We Work page. If more than one platform is involved, our AI adoption consulting engagements run the same pattern across all of them.

Has this worked at a real firm?

Yes. At Forvis Mazars, regional, CPE-accredited, instructor-led workshops trained 300+ professionals on Microsoft Copilot. The program standardized account planning firm-wide and refreshed the firm's operating model to embed guidance on when and how to use AI. The lesson we took from it: people adopt Copilot when training is built around the work they are accountable for, delivered in the tools they already use, and backed by clear rules. The full story is on our case studies page.

Should we cut Copilot licenses instead?

Maybe later, but not first. Cutting seats before anyone has been trained tells you nothing, because you do not know whether low usage came from low value or from low enablement. Run the program, then look at the per-user report again. After real, role-based training, the people who still never use Copilot are a much more honest signal, and reassigning those seats becomes a decision based on evidence rather than frustration.

Where do we start this week?

Open the Copilot usage report, export the per-user table, and sort by last activity date. Pick the two teams with the most enabled-but-inactive people and ask each lead one question: what is the email, meeting, or spreadsheet task your team dreads every week? That answer is your first workflow build.

If you would rather not do it alone, book an AI briefing. We will look at your usage data and your teams and map the fastest path from assigned seats to daily use.