A lot of companies bought the licences first and told managers second. Staff got a login and a short course. Their manager got an email asking them to encourage adoption. Then everyone waited for the usage dashboard to move.
Gallup found that only 28 percent of employees in organisations that have begun implementing AI strongly agree their manager actively supports the team's use of it. Employees who do strongly agree are 2.1 times as likely to use AI a few times a week or more. That gap sits with the manager, so the manager is who I would train first.
What a manager has to be able to do
A manager needs two things before the team will move. They need to know which of the team's tasks AI can help with this month. They need to use it on their own work where the team can see it. Gallup describes the manager's job here as modelling use and connecting it to the work people actually do. A manager who has never drafted anything with a model can do neither.
They also have to answer the awkward questions. What can go into the tool? Who checks the output before it reaches a client? What happens when someone gets it wrong? Does using it count against you in a performance review? If the manager shrugs, the team keeps using AI in a private window, or stops.
How the session should run
Train managers on their own work first. In our AI training workshop people map high value, low risk use cases for their own roles and get hands-on reps with tools they can keep. They leave with a 90 day plan to put AI into real workflows.
For managers I add one job to that plan. They pick a task the whole team does and run it with AI in front of the team. Then they write down the check a person still does before the work goes out. What a single day should leave people able to do is in what a corporate AI training day should produce.
Keep managers from the same function together. A finance lead and a marketing lead have different tasks and different risks, and a mixed room drifts back to generic demos. Role-based AI training makes the same case for staff cohorts.
What to look for a month later
Ask each manager one question after four weeks. Which task does the team now do differently, and who checks it? A named task and a named checker means the training worked. A shrug means you paid for awareness.
If your rules on data are still unclear, fix them first. Managers will not champion a tool they are unsure they are allowed to use. AI governance for non-lawyers is a sensible place to start.
If you want to start with your managers, tell me how many you have and which tools your staff may use. We will build the day around the tasks their teams already do.
