Best AI prompts for managers

Why managers use AI for the writing, not the managing

AI prompts for managers work because the hardest parts of people management arrive as writing tasks: performance reviews, feedback that has to land right, hiring scorecards, onboarding plans, and the follow-up email after every difficult conversation. ChatGPT and Claude can't run your team. What they can do is take the two hours of drafting wrapped around each people decision and shrink it to twenty minutes of editing.

One distinction before anything else. This page covers the people side of managing: leading a team, developing direct reports, making hiring calls. If your week is mostly project lifecycle work, the status updates and stakeholder communication and risk tracking, the project managers page collects those prompts instead. Plenty of managers need both pages. They're different jobs that happen to share a title.

Hiring: the manager's side of the process

Hiring prompts help most at the two points where you, not the recruiter, own the writing: defining the role and evaluating the people. The job description prompt turns your mental model of the role into a draft you can hand to recruiting, and it forces a question worth answering anyway: what does success in the first six months actually look like? The interview question prompt builds a structured question set for each competency, so two candidates for the same role get comparable interviews instead of whatever occurred to you that morning.

When a former report asks for a reference, the reference letter prompt structures the letter around specific accomplishments. Give it one real anecdote. The anecdote is the only part anyone remembers.

HR and recruiting run their own side of this pipeline, and the HR managers page covers the policy and process end. The prompts here serve the decisions that stay with you as the hiring manager: who to interview, what to ask them, whether to extend the offer.

Onboarding without reinventing the checklist

A new hire's first two weeks usually live in the manager's head, which is why every onboarding repeats the same scramble. The onboarding checklist prompt gets it out of your head and into a document: accounts, intro meetings, first tasks, and what good looks like at thirty, sixty, and ninety days. The SOP prompt does the same for your team's recurring processes, and a documented process is the difference between delegating a task and delegating a mess.

The introduction email prompt handles the note that introduces your new hire to the team and the partners they'll work with. It's a small message that sets up their first ten conversations.

Performance reviews, feedback, and 1:1s

Six direct reports means twelve review drafts a year, plus promotion cases, plus all the feedback in between. This is the pile that pushes managers toward AI first, and it's where the division of labor matters most: you bring the evidence and the judgment, the model brings the structure. The performance review prompt turns scattered notes into an organized draft, but it cannot remember what your report shipped in March. Keep a running note per person and review season stops being archaeology.

For growth conversations, the goal setting prompt converts "get better at stakeholder communication" into something concrete enough to revisit next quarter. The meeting agenda prompt adapts cleanly to 1:1s, where the agenda should belong to the report rather than the manager. And when you want a read on team health beyond your own impressions, the employee survey prompt drafts pulse-check questions people can answer honestly.

One rule sits above all of these. Describe people hypothetically, never by name: "a senior engineer, strong technically, tends to dominate discussions" gives the model everything it needs. Names, ratings, and salary details don't belong in a public AI tool. The specifics go back in after the draft leaves the chat window.

Difficult conversations

No prompt has the difficult conversation for you, and you should distrust any page that claims otherwise. What AI does well is the preparation: structuring your opening lines, anticipating reactions, taking the heat out of a first draft. The tone adjustment prompt exists for the message you wrote angry; it returns the version that says the same thing without the edge. The apology email prompt covers the moments when the mistake was yours, a message managers write more often than anyone warns them.

Logistics carry weight here too. The meeting request email prompt sets up the conversation without creating a scary vacuum of context, and the follow-up email prompt puts what you agreed on in writing while it's still fresh. The written follow-up is the step most managers skip, and it's the step that makes the conversation actually change anything.

The team's weekly rhythm

Recurring team communication is the most automatable writing a manager owns, because the format never changes and only the details rotate. The daily standup prompt keeps async updates specific enough to be worth reading. The meeting summary prompt turns team meeting notes into decisions with owners before memories diverge. The internal memo prompt handles announcements, the reorg note or the process change, where being read matters more than being thorough.

Two smaller ones earn their spot. The reminder email prompt chases timesheets and survey responses without making you sound like a nag, and the team building activity prompt saves the next offsite from another round of two truths and a lie.

Delegation and decisions

Delegation usually fails at the handoff, when a task that lives fully formed in your head gets described in one sentence. The task prioritization prompt helps you decide what to hand off at all: sort the list by what only you can do, then delegate from the rest. For coaching moments, the explain concept prompt produces the same explanation at three depths, useful when a junior report needs the scaffold version and a senior one needs the edge cases.

People decisions deserve more structure than a gut call made at 5pm. The decision matrix prompt makes your criteria explicit before you score the options, which matters for promotion cases and team structure changes where you'll need to defend the reasoning later. The pros and cons prompt is its lighter cousin for smaller calls. The brainstorming prompt earns its keep in retros and planning sessions, and the procrastination buster prompt is for the task you've dodged all week, which for most managers is a difficult conversation wearing a to-do item's clothes.

Managing up

Your team's work is invisible to the level above you unless somebody writes it down, and that somebody is you. The executive summary prompt compresses a quarter of team output into the one page your director will actually read. The summarize long document prompt works the other direction, reducing the forty-page strategy deck to the two pages your team needs. The weekly review prompt gives Friday afternoon a structure and quietly builds the evidence file that makes review season easier for everyone, including you.

The out of office prompt is the smallest prompt on this page with the clearest signal. A manager whose auto-reply names a delegate with real authority has been doing delegation right all year.

Which model for which management task

Model choice matters less than the specificity of your brief, but for people-management writing the differences are real. The full breakdown is in the ChatGPT vs Claude vs Gemini comparison; the short version for managers:

Gemini is the practical pick if your company runs on Google Workspace, since it drafts inside the Docs and Gmail your team already lives in.

What AI can't do for a people manager

It can't notice that your strongest engineer went quiet in standups three weeks ago. It can't tell you whether a performance problem is skill or motivation, and it can't feel the difference between a report who needs pushing and one who needs protecting. It has never met your team, and every draft it produces is written for a team it has never met.

The failure mode worth naming: a manager who sends AI-drafted feedback unedited will eventually get caught by its blandness. People can tell when a message about them wasn't really written for them, and with feedback, that recognition costs you trust you can't easily buy back. Use the draft to start. The sentence that makes it land has to be yours.

Where to start

Pick the writing task from last week that took you the longest. For most managers that's a review draft or a difficult email, and both are covered above. Run the prompt once with real context, hypothetically described, and compare the editing time against your usual blank-page time. That one comparison tells you more than any tool roundup.

When you're ready to improve the outputs on purpose, the five variables in the beginners guide transfer directly to management writing. Both OpenAI's prompt engineering guide and Anthropic's land on the same principle: the model is only as specific as your brief. "Write a performance review" produces boilerplate. "Draft a review for a mid-level designer, strong on craft, growing on stakeholder communication, with these three examples" produces something you'd almost sign.

The managers who still use these prompts a year in share one habit, and it isn't prompting skill. They keep a running note per direct report. The prompt writes the review; the note is what makes it true.