Answer
AnswerWe rolled out Copilot or ChatGPT Enterprise and few people use it. What now?
Low use after an AI rollout is rarely fixed by another announcement: show each role two or three tasks from its own work that the tool does well, have people practice on those tasks, and measure use by team.
A license gives access. It does not tell an accountant or a store manager what to do differently on Monday. Most generic introductions explain what the tool is, and people already know that. What is missing is the link to their own tasks. Start from the work: pick the recurring tasks in each role where the tool clearly helps, teach those with real examples, and let people practice on their own material. Then track use by team, so you can see where it took.
At a glance
- Usual cause
- People do not know which of their own tasks the tool helps with.
- What works
- Role-specific tasks, practiced on real work.
- What does not
- One generic webinar for everyone.
- What to measure
- Use by team some weeks later, not attendance.
What are the steps to raise use of an AI tool people already have?
Find out who uses it and for what
Pull usage by team from the tool’s admin console and ask a few regular users what they use it for. Their tasks are your first examples.
Pick tasks per role
For each role choose two or three recurring tasks where the tool saves real time. A short list people recognize beats a long one they do not.
Teach with their own material
Have people run the task on a document or dataset from their job. Practice on real work is what carries over to the next day.
Cover the rules at the same time
Say what data may go into the tool and what must not. Uncertainty about the rules is one reason careful people hold back.
Measure and repeat
Check use by team after a few weeks, share what worked in one team with the next, and add tasks as the tool changes.
Why is adoption low even when the tool is good?
Because access and habit are different things. In a two-year trial of an AI tutor that was one click away, 96 percent of students tried it at least once, and the median student used it in only 17 percent of the exercise sessions in which they made a mistake (Oreopoulos and Low, NBER, 2026). Availability did not produce use. A general-purpose assistant at work also waits to be asked, and people who do not know what to ask do not ask.
Where does Scholé fit after a rollout?
Scholé generates a lesson for each person from their role, tools and tasks, so the examples are their own. Its browser extension, Learn Anywhere, gives the learner a real task inside the tool and guides them through it from a side panel. For the organization, it reports adoption by team.
Scholé is the training, not the assistant: you keep Copilot, ChatGPT, Gemini or Claude, and Scholé teaches people to use them.
Where Scholé fits
- You have licensed an AI tool and want training built around each role’s real tasks.
- You want people to practice inside the tool itself, with guidance, and adoption reported by team.
Where Scholé is not the answer
- The tool itself is the problem, for example it cannot reach the systems people work in. Training does not fix a missing integration.
- You have not decided what staff may put into the tool. Settle the policy first, because training cannot answer a question the organization has not.
- You want a general chatbot for staff. That is what Copilot, ChatGPT, Gemini and Claude are, and Scholé teaches their use without replacing them.
More on this site: Scholé for teams, Learn anywhere, Scholé vs ChatGPT.