How widespread is the problem?
Microsoft has reported that a significant proportion of Copilot licences across enterprise customers are either unused or used only once in the first month after activation. The reasons vary, but the pattern is consistent: organisations buy licences as part of a broader Microsoft 365 agreement, or in response to a board directive to "do something about AI," and then discover that adoption doesn't follow automatically from access.
If your organisation has Copilot licences, there is a reasonable chance that a meaningful percentage of them are sitting idle right now. The cost adds up quickly — at current pricing, an unused licence represents hundreds of pounds per user per year in wasted spend.
Why licences go unused
No training at the point of rollout. The most common cause. Licences are activated, an email goes out, and people are expected to figure it out themselves. Most don't. The tools aren't intuitive enough for self-directed adoption, and without someone showing people specifically how Copilot applies to their actual job, the initial curiosity fades within a week.
No clear use cases communicated. "You now have access to AI" is not a use case. People need to know what problem it solves for them, in their role, in their workflow. Generic capability statements don't create behaviour change. Role-specific examples do.
Fear of doing it wrong. A surprising number of people avoid using AI tools because they're worried about the output — about sharing something generated by AI that turns out to be wrong, or about being seen to rely on a tool they don't fully understand. Without guidance on appropriate use and quality checking, this anxiety leads to avoidance.
No visible leadership adoption. If senior people in the organisation aren't visibly using Copilot, it signals to everyone else that it isn't really expected. Culture follows leadership. When managers start using Copilot in meetings and referencing it in their work, adoption accelerates across their teams.
The wrong licences for the actual use cases. Not all Copilot licences are equal. Microsoft 365 Copilot, Copilot for Sales, Copilot Studio, and the various other variants have different capabilities and different appropriate use cases. Organisations sometimes find they've licensed a product that doesn't match what their teams actually need.
How to recover the value
The first step is understanding what you actually have. A licence audit — going through which licences are assigned, to whom, whether they're being used, and what the usage looks like — gives you the baseline. Microsoft's admin centre provides usage data, though it takes some interpretation to turn into a clear picture.
Once you know what you have, the question is whether the issue is training, communication, use case clarity, or something structural. Usually it's a combination of the first three.
Targeted training — not a generic AI overview but a session built around the specific roles and workflows of the people in the room — is consistently the highest-impact intervention. When someone sees exactly how Copilot can help them draft a document they write every week, or summarise a meeting they attend every Monday, adoption follows.
For larger organisations, identifying internal champions — people who are already using Copilot effectively and are willing to share what they've found — can accelerate adoption without requiring central resource. Peer-to-peer learning tends to land better than top-down instruction.
The audit as a starting point
If you're not sure where your organisation stands, a Copilot licence audit is a useful first step. It identifies unused or underused licences, flags mismatches between licence type and use case, and produces a clear picture of where the value is currently being left on the table.
It also tends to surface broader questions about AI readiness — about what training people actually need, and what the realistic adoption path looks like for different parts of the business.
Unused licences aren't a sign that AI doesn't work. They're usually a sign that the rollout stopped at access and never got to adoption. The gap between having the tool and using it effectively is where most of the value is lost — and where the right training makes the biggest difference.