Delta

Practical guides6 min read

Copilot adoption: why the licences sit unused, and a 30-day fix

Copilot licences often sit idle because the work around them never changed. A practical 30-day plan to get Copilot adoption moving in your team.

Newan VinthusaFounder, Delta Camp
Illustration of a tray holding a row of folders, with one folder pulled out and placed beside a small timer.

The quickest way to see your copilot adoption problem is to open the Copilot usage report and count the licensed people who haven't touched it lately. Microsoft says the report can show the number of inactive users over a time range, for example people who didn't use Copilot in the past 28 days. You don't need a survey to find out whether your licences are sitting idle.

Whatever number you see, it's rarely a laziness number. Those people have a licence, a full inbox and a week that already runs a certain way. Copilot sits in a side window next to that week and never becomes part of it.

This guide is a 30-day plan for changing that, written for a manager who owns a team's results but not the Microsoft 365 admin settings. You need read access to the usage report (or a colleague who has it), one team, and some protected time each week.

Why the licences sit unused

At Delta Camp we treat Copilot as almost never a training problem. The tool isn't fitted into the day people already have, so it stays a curiosity. Microsoft's own research points the same way: organisational factors such as culture, manager support and talent practices account for about two-thirds (67%) of reported AI impact, against 32% for individual mindset and behaviour (knowledge workers using AI, across 10 markets).

The gap shows up in Asia Pacific too. 78% of APAC employees use AI at least weekly, yet only 57% say their companies are redesigning workflows to integrate AI effectively (employees across nine APAC markets). People are experimenting on their own while the work around them stays exactly as it was.

Two more findings make it worse. Only 51% of non-managers feel they have the learning resources they need, compared with 72% of senior executives (a global workforce survey), so the people doing the repeat work are the least supported.

The second finding is about roles. In ASEAN, the ILO finds that workplace use of generative AI remains comparatively limited in many clerical and administrative occupations, despite their high exposure. Those are often the roles with the most to gain.

A 30-day copilot adoption plan

Forrester describes enterprise CIOs and CDOs testing Copilot in targeted scenarios before committing to a broader rollout. Do the same at team size: one team, three tasks, one month.

Week one: read the report and pick one team

Export the usage report and split the inactive users by team. Pick one team whose manager will back the experiment, because manager support is one of the organisational factors in that Microsoft research. Write down the team's inactive count today so you have a starting point.

Week two: audit the work, then choose three tasks

Ask each person to list the tasks they repeat, with a rough time against each. Most will live in Outlook, Teams, Excel, Word, PowerPoint or SharePoint, which is where Copilot already sits. Then sort each task into what AI does alone, what it does with you, and what stays human, the method I explain in Audit before tools.

Pick three tasks, not fifteen. Start with the most frequent ones, because frequency beats size almost every time. A weekly status email is a better first target than a quarterly report.

Week three: put the three tasks into the calendar

A task that depends on someone remembering to try Copilot will not survive a busy Thursday. Block the slot where the task normally happens, write the prompt into a shared document, and name one person who owns each task.

Anything a customer will see goes past a human first, and someone writes down how the output gets checked. (This is the annoying step. Skip it and you meet the four-hour tax.)

Week four: re-read the report and measure what stopped

Pull the usage report again and compare the team's inactive count with the number you wrote down. That tells you whether people are touching the tool, but not whether the work changed. The better question is which task stopped being done by hand.

Decide where the freed time goes before you celebrate, or it leaks into the inbox. I explain why in the capture gap.

I can't promise the inactive number drops to zero in a month. It is the cheapest number to watch, though, and a team with three tasks running on Copilot is a better pitch to the next team than any demo.

What I'd skip

  • Skip the company-wide launch webinar, because a demo doesn't change anyone's Tuesday.
  • Skip prompt contests, since they reward clever prompts instead of repeated work.
  • Don't wait for a perfect policy document before one team starts; write the rules for the three tasks and widen them later.
  • Stop counting licences as the success measure, and count the tasks that now run without you doing them by hand.

Where Delta Camp's Copilot programme fits

If you'd rather do this alongside others in a live room, Delta Camp's AI training for Copilot is the Microsoft Copilot Champion programme. It runs for six weeks with one live session a week, built on your own work. You need a Microsoft 365 Copilot licence, but no admin rights and no code (and it is Delta Camp's own programme, not official Microsoft training).

To see where your team stands before you start, the free AI Maturity Audit takes about three minutes for an individual and about six for a business.

One thing to do this week

Open the usage report, write down this month's inactive count, and pick the one team you'll work with. That's the whole job for the week. Everything else in the plan starts from that number.

Frequently asked questions

How do I check whether our Copilot licences are being used?

Microsoft's Copilot usage report shows how many licensed users have been inactive over a time range you choose. If you don't have admin access, ask whoever manages Microsoft 365 to share it by team.

Why isn't anyone using Microsoft Copilot at work?

Usually because Copilot was switched on but never fitted into the tasks people already do each week. Without a specific task, a named owner and a manager who backs it, it stays a side window that nobody opens.

Is low Copilot adoption a training problem?

Rarely. A one-off demo or video course doesn't change a week that is already full. What helps is picking a few repeat tasks and building Copilot into them with the team.

Should we roll Copilot out to the whole company at once?

I'd start with one team and three tasks, then widen once something is visibly running. A small win is easier to copy than a company-wide launch, and it tells you what your rules need to cover.

Do I need a Copilot licence to join the Copilot Champion programme?

Yes, you need a Microsoft 365 Copilot licence at work. The free Copilot in Edge or Bing isn't the same thing. You don't need admin rights or any coding.

Where the numbers come from

Keep reading

Illustration of a folder with a sheet of paper half pulled out, a pen resting across it and a small timer standing beside them.

Practical guides5 min read

An AI readiness checklist for your team: 12 questions to answer before you buy anything

A practical AI readiness checklist for your team: 12 questions on data, people, risk and value to answer before you spend on any tool.

Read the post
Illustration of a desk with orange, green and grey sticky notes sorted into three columns, and a closed laptop pushed to one side.

Practical guides6 min read

Audit before tools: how to decide what to automate first

A 30-minute exercise for deciding which tasks to automate first: list your week, time it, and sort every task into three buckets before you buy an AI tool.

Read the post