Many companies have now bought AI tools for their staff. A few months later, the pattern is familiar. A handful of people use them every day, most people tried them twice, and nobody is sure what the licences are delivering.

The research matches what I see. In a WRITER survey of 2,400 executives and employees published in April 2026, 79% of organisations reported challenges with AI adoption. A Gartner survey in March 2026 found 86% of managers struggle to drive AI adoption in their teams. The tools work. The adoption plan is usually missing.

Why people stop using AI

When I train colleagues, the same five reasons come up.

  • No clear task. People are told to use AI, but nobody says for what.
  • Weak first results. A short prompt returns a generic draft, and people decide the tool is not useful.
  • No company knowledge. The tool knows the world but not your services, clients or style.
  • Unclear rules. People are not sure what data is allowed, so they hold back.
  • Managers not leading. If the manager never uses it, the team reads AI as optional.

None of these is a technology problem. Each one has a practical fix.

A 30-day plan for one team

Start small. One team, one month, three tasks that repeat every week.

  • Week 0. Pick the team and three tasks. Social posts, meeting summaries and first drafts of standard emails are good places to start.
  • Week 1. Write a model brief for each task: the goal, the reader, the facts, the limits and an example of good output. Test the briefs together.
  • Week 2. Save the briefs that work in a shared place, with a strong sample of each.
  • Week 3. Ask each person to share one result in the team meeting. Seeing a colleague's good draft does more than any demo.
  • Week 4. Count the tasks done with AI, the time saved and what the time went to.

The last step matters more than it looks. Gartner found only 7% of organisations guide staff on what to do with the time AI saves. If the saved hour has no purpose, the habit fades.

Managers go first

The most reliable signal I have seen is simple. When a manager shows one real AI result at a team meeting, the team starts trying. When a manager never mentions it, the licences sit unused. A manager does not need to be the best user on the team. They need to show that using it well is part of the job.

Keep people in charge of the result

Adoption does not mean publishing whatever AI produces. The teams that keep going have a named person who approves each type of output. That is what makes the work trustworthy, and it is what gives people the confidence to keep using the tools.

Pick one team and three tasks this week. I would love to hear how the first month goes.

Download the two-page guide: Get the crew working (PDF), Guide 4 in the Director's Chair series. It is free to share with your team.