Lenovo: 36% higher daily copilot use came from training

Lenovo compared two groups of Microsoft 365 Copilot users. One group received licenses only. The other also received role-specific training, early manager involvement and continued feedback. The supported group reached 91% daily use against 67% for the license-only group. That gap is what “Lenovo: 36% higher daily Copilot use from its adoption model” refers to, and it is reported by Lenovo on news.lenovo.com.

Many leaders still assume that once people have access to an AI tool, usage follows. Lenovo’s own early deployments showed that initial activity did not always last. The comparison puts a number on the difference between handing out licenses and running an adoption model.

This article walks through what Lenovo did, why it worked, and how to plan the same thing in a company that does not have Lenovo’s scale.

Lenovo: 36% higher daily Copilot use from its adoption model, in numbers

The headline figure is daily use. The more useful figures sit underneath it.

Measure	License-only cohort	Supported cohort
Daily use	67%	91%
Time saved per person per week	1.8 hours	3.5 hours
Satisfaction (out of five)	3.67	4.27
Issue rate	33%	18%

The time saved matters most to me. A supported employee saved 3.5 hours a week. A license-only employee saved 1.8. Both cohorts had the same tool and paid the same license cost. The supported group got close to double the return on it.

Run that through your own numbers. Take the number of Copilot seats you plan to buy and multiply by the 1.7 hour weekly difference. That is the capacity you leave unused if you skip the training. For most SMEs that figure is larger than the training budget.

The issue rate is worth a second look too. It fell from 33% to 18%. Fewer people got stuck, which means fewer tickets, fewer workarounds and fewer people quietly giving up on the tool.

Why licenses alone stall Copilot adoption

The usual rollout plan has two steps. Buy licenses. Announce the launch. Sometimes a generic demo is added in between.

That plan produces early activity. People open the tool, try a few prompts, summarise an email or two. Then usage drifts. Lenovo observed this pattern in its own early deployments, which is why it built a different model.

My reading is simple. Adoption comes with awareness. People use an AI tool once they see where it fits in their own tasks. A generic demo shows what the tool can do in general. It rarely shows a procurement analyst, a finance controller or a sales manager what it does for the report they write every Monday.

Without that link to real work, the tool stays an optional extra. Optional extras get dropped when the week gets busy.

How role-specific training raised daily Copilot use at Lenovo

Lenovo accepted that generic training could not cover every role. It moved to formats built around the employee’s own work.

Prompt-a-thons on real tasks

Lenovo ran hands-on prompt-a-thons. Employees brought their actual work and applied Copilot to it in the session. They left with prompts that solved a problem they had that week.

This is the part I would copy first. A prompt-a-thon is cheap to run. It needs a room, a facilitator who knows the tool, and participants who bring a real document, a real meeting or a real dataset. The output is a set of working use cases per team.

Persona-based enablement

For the rollout of its Knowledge Super Agent, Lenovo used persona-based enablement. Training was designed around types of users and what each type needs to get done. A sales persona and an engineering persona get different examples, different prompts and different success criteria.

An SME can do a lighter version. List the five or six roles that will use the tool most. For each one, write down three recurring tasks where Copilot can save time. Build the training around those tasks.

A champion network

Lenovo also set up a champion network. Champions are colleagues who use the tool well and help others in their team. They answer the small questions that never reach a helpdesk.

Champions matter because most AI questions are local. Someone wants to know how a colleague drafted that client summary. The fastest answer comes from the person at the next desk.

Why managers belong in the Copilot rollout before launch

Lenovo’s early rollouts left managers little time to prepare. Employees look to their manager to explain why a change matters. When the manager has not been briefed, that explanation does not happen.

Lenovo now brings managers in before training and before launch. Three things follow from that.

1. Managers can explain the purpose. They connect the tool to the team’s goals and workload.
2. Managers spot resistance early. They know who is worried, who is overloaded and who will need more support.
3. Managers surface role-specific needs. They know which tasks in their team are repetitive and where time is lost.

AI training builds confidence quickly. Real usage takes longer and usually needs a nudge from the top. The manager is that nudge at team level. A director’s announcement sets the direction. The manager’s weekly follow-up decides whether the habit forms.

In practice this means a short manager session one or two weeks before the team training. Cover what the tool does, which team tasks it targets, and what the manager is expected to do afterwards. Keep it practical. Managers need to use the tool themselves before they can ask their team to.

How Lenovo kept Copilot usage going after launch

Lenovo learned that initial activity did not guarantee lasting use. So it measured usage over time.

Adoption dashboards tracked usage after the launch. Feedback channels surfaced barriers that the team could then fix. The issue rate falling from 33% to 18% is the visible result of that loop.

Measurement turns adoption into something you can manage. Without it, a rollout gets judged on anecdotes. With it, you can see which teams have stalled, which use cases took hold and where a second session is needed.

For an SME, the Microsoft 365 admin reporting gives a first view of Copilot activity. Pair it with a short monthly pulse question to users: where did the tool help, and where did it get in the way. That is enough to steer.

What the Lenovo adoption model means for an SME budget

The practical lesson from Lenovo’s 36% higher daily Copilot use is about where money goes. Training belongs in the AI budget next to licenses. Both cohorts had the same tool. The return was very different.

AI amplifies what already exists. Licenses dropped into a team that does not know where the tool fits amplify very little. Licenses dropped into a team that has mapped its tasks, briefed its managers and measured its usage amplify real work.

Here is the sequence I recommend before buying more seats.

1. Pick the roles. Start with the teams where repetitive writing, summarising or analysis takes the most hours.
2. Map the tasks. For each role, list the recurring tasks where Copilot can help. Start from the pain in the work, then look at the tool.
3. Brief the managers first. Give them time to use the tool and prepare the conversation with their team.
4. Train on real work. Run prompt-a-thons where people bring their own documents and meetings.
5. Name champions. One per team is enough to start.
6. Measure for three months. Track daily use, time saved and the problems people report. Fix the barriers you find.
7. Expand the licenses only after the first teams show steady use.

This sequence takes a few weeks of preparation. It is the difference between 67% and 91% daily use in Lenovo’s case, and between 1.8 and 3.5 hours saved per person each week.

The decision to make before the licenses arrive

Training is the success building block. Lenovo’s comparison makes that easy to argue in a budget meeting.

Before signing the next license order, settle three questions. Who trains whom. On which tasks. How usage gets measured. If those answers exist, the licenses will be used. If they do not, expect the activity curve Lenovo saw in its early deployments.

The tool recommends and drafts. Your people decide what to do with the output. The adoption model is what gets them to the point where they trust it enough to try.

FAQ

What is Lenovo: 36% higher daily Copilot use from its adoption model?

It is the result of an internal Lenovo comparison on Microsoft 365 Copilot. A cohort with role-specific training, early manager involvement and continued feedback reached 91% daily use. A license-only cohort reached 67%. The difference is 36%, as reported on news.lenovo.com.

How much time did Lenovo employees save with Copilot?

Supported employees saved 3.5 hours per person each week. Employees with licenses only saved 1.8 hours. Satisfaction also rose from 3.67 to 4.27 out of five, and the issue rate fell from 33% to 18%.

What is a prompt-a-thon?

A prompt-a-thon is a hands-on session where employees apply an AI tool to their own work. Lenovo used them in place of generic demonstrations so people could see where Copilot fits in their real tasks.

Why should managers be involved before an AI rollout?

Employees look to their manager to explain why a change matters. Lenovo found that early rollouts left managers little time to prepare. It now involves them before training and launch so they can explain the purpose, spot resistance and identify role-specific needs.

Can a small company copy the Lenovo Copilot adoption model?

Yes, in a lighter form. Pick a few priority roles, map their recurring tasks, brief managers first, train on real work, name one champion per team and track usage for three months before expanding licenses.