What to tell your team when they ask if AI is coming for their job.

Sooner or later someone on your team will ask it, probably at the end of a meeting about something else. Is AI going to take my job? How you answer that question will shape how your people use AI for the next year, so it is worth deciding what you will say before you are asked.

Why the question deserves a real answer.

Most leaders reach for reassurance. "Nobody is losing their job" feels kind, and it ends the awkward moment. The trouble is that people can tell when a promise is bigger than the person making it. If you can't guarantee every role for the next five years, your team knows you can't, and a guarantee you can't keep makes everything else you say about AI sound like spin.

The opposite mistake is the shrug: "Things are changing, we'll all have to adapt." That sounds honest, but it hands the worry back to the person who raised it with nothing to hold onto. They leave the room with the same fear and less trust.

Somewhere between those two is an answer you can actually stand behind. It names what is changing, what isn't, and what you are committing to as an employer. It also admits what you don't know.

Start with what is actually changing.

Be specific about the work, because the work is what's moving. In most organizations we see, AI is taking over pieces of tasks: the first draft of a proposal, the summary of a long email thread, sorting an intake form into the right queue, reconciling a stack of invoices against purchase orders. Those pieces used to fill hours. Some of those hours will go away.

Say that plainly. If the person who processes permit applications at a town office spends a third of the day retyping information from PDFs, and a tool can now do the retyping, that part of the job will shrink. Pretending otherwise insults the people who can already see it happening on their own screens.

If you can't guarantee every role for five years, a guarantee is the one thing you shouldn't offer.

Then be just as specific about what goes with the task. Someone still decides whether the application is complete. Someone still calls the applicant when a detail doesn't add up, and someone answers for the decision when it goes to the board. AI changes how much time the clerical part takes. It hasn't changed who the organization trusts to make the call.

Say what isn't changing.

People need to hear what stays fixed, and it should be things you control. The relationships with clients and residents, the judgment calls, the knowledge of how this particular place works, the reputation that comes with someone's name on a piece of work. A model can draft the letter to a family whose claim was denied. It can't be the person the family calls back, and you are not planning to make it one.

It also helps to describe how decisions about roles get made in your organization, because that process isn't changing either. If headcount decisions have always gone through a budget cycle, a conversation with the department head and some notice, say so. Much of the fear around AI comes from imagining that a decision will arrive overnight from somewhere above. Showing people the ordinary path a decision travels makes it smaller.

Make commitments you can keep.

This is the part most leaders skip, and it is the part people remember. Commitments turn a vague reassurance into something your team can check you against later.

A few that most organizations can make honestly:

• When AI frees up time in a role, we will talk with the person in that role about where the time goes before anyone else decides.
• We will train people on these tools during work hours, at our expense, before we expect them to use the tools well.
• Nobody will be judged on AI skill they haven't been given the chance to learn.
• If we ever do plan to change a role significantly, the person in it will hear it from their manager first, with time to prepare.

Each of these describes a process, and a process is something a leader can actually deliver, even in a year when nobody can say exactly how every job will look. If you can't sign up to one of these, leave it out. A short list you keep is worth more than a long one you drop.

Keep the conversation going.

One good answer in a meeting won't settle the question. People will ask again when a new tool shows up, when a colleague leaves and isn't replaced, or when a news story makes the rounds. Treat it as a recurring topic and give it a regular slot. A few minutes in a monthly team meeting, where people can say what has changed in their work and what worries them, will catch rumors early.

Managers carry most of this, so give them the same talking points you use and let them practice. A department head at a clinic or a school district office will get the question in private, often from the person most anxious about it. They need to know what the organization has committed to, what it hasn't, and when to say "I don't know, and I'll find out."

Talking points for the conversation

Put these in your own words, but keep the order.

1. Thank them for asking. Say it's a fair question and you'd rather hear it than have people worry in silence.
2. Name the tasks in their role that AI is already changing or likely to change. Be specific, and ask them to add to the list.
3. Name what isn't changing: the judgment, the relationships and the accountability that stay with people.
4. Explain how decisions about roles get made here, and that this process hasn't changed.
5. State your commitments, only the ones you will keep, and say where they're written down.
6. Admit what you don't know. You can't predict every tool or every year, and saying so is more believable than a guarantee.
7. Offer one next step for them: a training session, time to try a tool on their own work, or a follow-up conversation in a month.
8. Close by telling them who to come to with the next question, and mean it.