How to Tell Employees About AI Changes: What to Say and What Not to Say

Most managers will have some version of this conversation in the next year. A team learns that a new AI tool is coming or a process is being redesigned, and the first question in the room is rarely about the tool. People want to know what it means for their jobs.

They arrive worried. In Gallup’s August 2026 poll, 27% of US workers said they worry technology could soon make their jobs obsolete, a record and up 7 points in a year. Yet only one in four employees say their organization has communicated a clear plan for AI. Rumors grow in the space between that worry and the missing information, and a well-run conversation closes it.

Why does the way you tell employees about AI matter so much?

The message travels through managers, and managers decide how it is heard. A Gallup study of nearly 30,000 worker observations from 2023 to early 2026 found that managers shape how workers interpret AI, and that broad communications about AI strategy are unlikely to reach the workers who have the fewest ways to reassure themselves.

Getting it wrong shows up in the work. In the same Gallup data, workers with at least some displacement concern scored lower on engagement and were 4.1 points more likely to be actively looking for another job. The Conference Board reports that workers who feel threatened by AI score more than 25 percentage points lower on engagement, effort and intent to stay.

When should you tell employees about AI changes?

Earlier than feels comfortable, but not before you have something concrete to say. If you wait until every decision is final, you lose the chance to involve the people who know the work best. SHRM’s guidance on engaging employees in AI without triggering fear advises explaining early why the change is happening and what roles will look like afterward, and inviting input before decisions are final.

A practical test is whether you can name the tasks. Before the first conversation, you should be able to say which parts of each role the change touches, which parts it leaves alone and what you do not know yet. A role-by-role view of AI exposure turns “AI will change things” into a list of tasks people recognize from their own week.

If the change could lead to layoffs, timing also has a legal side. The federal WARN Act generally requires employers with 100 or more employees to give at least 60 calendar days of advance written notice of a plant closing or a mass layoff affecting 50 or more employees at a single site, and some states add their own rules. This is general information, not legal advice, so bring employment counsel in before you set dates.

Who should hear about AI changes first?

The order matters more than most leaders expect. The leadership team agrees on the facts first. Managers then get a briefing with the role-level details and a written FAQ, and each team hears the news from its own manager on the same day as any company-wide note, so nobody learns about a change to their job from an all-hands slide.

Managers go first because employees trust people close to them. In Edelman’s 2025 flash poll on trust and AI, “someone like me” was on average two times more trusted than a CEO or government leader to tell the truth about AI, and Edelman urged employers to favor peer-to-peer communication over top-down missives. Managers also need the same facts as executives: BCG’s 2026 AI at Work survey of close to 12,000 people found that only a third of frontline employees say leadership’s communications about AI are clear, and only 28% see a strong connection between what leaders say and what the organization actually does.

Do not assume managers are ready. In Gallup’s tracking, only 36% of employees in organizations that have integrated AI strongly agree that their manager supports their team’s use of it. Rehearse the hardest questions with managers before they face their teams.

What should the first message about AI say?

Keep the first conversation short and specific. It should cover five points:

1. What is changing and why. Name the tasks and the business reason in plain words. Gallup recommends connecting AI to what the team will be able to do, not only to headcount efficiency.
2. What is not changing. People relax a little when they hear which parts of their work stay the same.
3. What is undecided, and when you will know. Give a date for the next update.
4. What support they will get. State the training hours and whether they fall on paid time.
5. How they can shape it. Offer a pilot or a working group, and a simple way to flag what will not work.

What does a good AI talk track sound like?

Here is a short example for a team whose work will change but whose roles are not being cut. Change every detail to match what is true for your team.

“I want to talk about something that affects our work, and I would rather you hear it from me than from a rumor. Starting in March, we will use an AI tool to draft first replies to routine billing questions.

This part is decided. The tool drafts, and one of us reviews every reply before it goes out. Complaints and any refund above the usual limit stay fully with us. Two things are still open: how we use the time this frees up, and whether our targets change. I will update you by the end of January, even if the answer is that we are still working on it.

Everyone gets four hours of training during work time in February. I also need a few volunteers to test the tool with me for two weeks and tell me what is wrong with it.

I know some of you are wondering about your jobs. I will not make promises I cannot keep, so here is what I can say: there are no planned reductions on this team, and if that ever changes, you will hear it from me first. What questions do you have?”

Use the last paragraph only if it is true. If it is not, see the section on shrinking roles below and plan individual conversations instead.

What questions will employees ask, and how should you answer?

Expect the same handful of questions on every team. Write honest answers into the FAQ so every manager gives the same one.

“Is my job going away?” Answer for the role, not the whole company, and say only what is true. If the role will shrink, do not answer in a group meeting; say you will speak with each affected person individually this week, and then do it.

“Will I get trained, and on whose time?” Give the hours, the dates and whether training happens during paid time. In BCG’s 2026 survey, only 36% of respondents felt they had received adequate upskilling.

“What happens to the time AI saves?” If you have a plan, describe it. If not, say you will build it with the team by a stated date.

“Do I have to use it?” Be clear about what is required and explain why. In Edelman’s poll, two in three AI distrusters in the UK and US felt AI was being forced on them.

“Can we push back?” Explain who made the decision and which parts are still open to input. Then show that input changes something, or people will stop offering it.

Which phrases backfire when you talk to employees about AI?

Some lines sound reassuring in a leadership meeting and land badly on a team. These four come up most often.

“No one will lose their job.” Say this only if you can guarantee it for the period people care about. IMD professor Amit Joshi writes that people often find it hard to believe senior leaders who insist AI investment will not lead to job losses, and that employees who think leadership is glossing over the impact are more likely to resist the change. Edelman found that two-thirds or more of respondents in developed countries believe business leaders may not be fully honest with employees about the impact of AI on jobs.

“AI will free you up for higher-value work.” This line only works with specifics. In BCG’s 2026 survey, 66% of frontline employees who use AI regularly said they receive limited or no guidance on what to do with the time they save. If you cannot name the higher-value work, people hear “less work, so fewer of us.”

“It is just a pilot” or “we are still exploring.” If the pilot is the first step of a rollout, say so, and if you are still exploring, give the date you will decide. Repeated for months, either line sounds like hiding something.

“This is about efficiency.” Efficiency may be one honest reason, but on its own it tells employees that the goal is fewer people. Connect the change to what the team will be able to do.

How should managers follow up after the first conversation?

Trust is built in the follow-up. Hold short one-on-ones within two weeks, because many people will not ask their real question in a group, and share each FAQ update with the whole team so people see that asking works.

Use more than one channel; SHRM’s guidance recommends repeating the message in town halls, manager one-on-ones and onboarding for new hires. Then report back on the dates you promised. If you said you would know about targets by the end of January, send a note in January even if the answer is “not yet, and this is why.”

What should you say when some roles really will shrink?

Say it plainly, to the people affected first, in private, because telling a whole team that “some roles may be affected” leaves everyone afraid. Meet each affected person, explain what is changing and why, give the timeline and lay out the real options, such as a move to another role, retraining or separation with support. Then put it in writing.

Be honest about the reason. If a reduction is driven by cost rather than AI, do not blame AI, and if AI is part of the reason, do not hide it. Since March 2025, New York’s WARN filing has asked whether technological innovation or automation contributed to a layoff, and as of May 2026 the law firm Hunton Andrews Kurth reported that none of the more than 160 companies that had filed notices since the change had attributed layoffs to AI or automation. Your public explanation, your filings and what managers tell their teams should all match.

Before anyone hears the word “reduction,” check the redeployment options. A person whose role is shrinking often has skills that fit a nearby role, and our guide to redeploying people before you replace them explains how to find those moves.

If you want the role-by-role facts before you talk to your teams, send JobRoute 10 job titles. We will return a free AI exposure sample within two business days, showing which tasks in each role AI can already do and which tools fit that work. You can request the free sample or book a 15-minute call with a founder.