“I’m the manager and I understand AI less than my juniors”

Written By Rafaela Heming

If you're a manager who doesn't understand AI as well as the people you lead, you are in the majority, the gap is narrower than it feels, and what your team needs from you is not technical fluency. It's direction: what we use it for, what we don't, and what "good" looks like. That takes a few hours of real hands-on use, not a course. And it can't be delegated, because the moment you outsource the judgment, you've outsourced the leadership.

I'm Rafaela Buena. I coach managers and business owners in Amsterdam through exactly this, one to one, and the first thing most of them say is some version of "please don't tell my team I booked this."

You're not imagining the gap

Gallup's Q1 2026 workforce study found that 50% of CHROs are not confident in their managers' ability to guide employees on using AI at work. The same research shows managers play a decisive role in whether AI-related change is perceived positively or negatively by employees.

A 2026 Acorn survey of over 1,200 professionals found a striking mismatch: 77% of executives believe their managers are prepared to guide AI skills development, but 91% of employees say their managers are not.

Fast Company's leadership coverage in 2026 put it plainly: senior leaders still hold experience and judgment, but younger employees increasingly hold the practical AI skills. Companies like Accenture, Target and Unilever now run reverse-mentoring programmes because of it.

So the feeling is accurate. What's inaccurate is what you conclude from it.

Why a manager who doesn't understand AI still leads the AI conversation

Your juniors are faster at the tool. That's real, and it's fine. What they usually don't have is the judgment about when to use it, what the output is worth, and what happens if it's wrong.

And there's a cost to getting this wrong in either direction. A 2026 Omni Calculator survey of 913 employees found 42% see managers who rely heavily on AI as less competent, and 36% trust their judgment less. Sixty-eight percent said AI-assisted messages feel less personal. Younger workers were the harshest critics.

Read that carefully. Your team doesn't want you to become an AI power user. They want you to know enough to set direction and stay human while doing it.

What you actually need to know

Less than you think. Four things.

  • What the tool is good at and bad at, from your own hands. Two or three hours of using it on your own tasks: a report, a difficult email, a decision you're weighing. Not a demo. Your work.

  • Where the risk lives. What data can't go in. Where it hallucinates. Which of your team's outputs need a human check before they leave the building.

  • What "good" looks like for your team. If you can't describe a well-used AI draft versus a lazy one, you can't set the standard. Do the checking habit yourself once and you'll be able to.

  • What you're not going to use it for. Gallup's research and the business.com 2026 survey both point to the same tension: people are deeply uncomfortable with AI in performance reviews, feedback, and personal communication. Decide your line and say it out loud.

This is most of what a 1-1 session with a manager covers. Not "learn AI." Two hours on your own tasks, quietly, until you can walk into a team meeting and set direction without bluffing.

What to delegate, and what not to

Delegate: building the setups, testing tools, finding the shortcuts, running the reverse-mentoring hour. Your fluent juniors will love it and it costs you nothing.

Don't delegate: deciding what AI is for on your team, the data rules, the quality standard, and the conversation about fear. Those are yours. If you hand them off, the Glassdoor data on "AI pushers" tells you exactly how it lands.

The sentence that changes the room

"I'm not as fast with this as some of you, and I'd like you to show me. But I'll be deciding what we use it for and how we check it." Said once, honestly, that sentence does more for your credibility than pretending ever will.

FAQ

Do managers need to be AI experts?

No. Research from Gallup and others shows the manager's role is to set direction and model attitude, not to be the fastest user. A few hours of hands-on use on your own work is enough to lead credibly.

Is it bad if my employees know more about AI than I do?

Only if you pretend otherwise. Reverse mentoring is now standard at large companies precisely because juniors often hold the practical skills. Being open about it builds trust; bluffing erodes it.

Should managers use AI to write feedback or reviews?

Most employees strongly dislike it. The business.com 2026 survey found 65% uncomfortable with AI-written reviews or feedback. Use AI to prepare, not to replace your own words on the things that matter.

If you'd like to close the gap privately, on your own tasks, before you have to set direction in public, that's what a 1-1 session is for. Two hours, no slides, nobody watching.

Sources

Rafaela Heming
Next

“Half my team uses AI daily and the other half never touches it”