AI for managers: 10 tasks you can hand over to AI

What you can and can't hand over to AI

AI handles drafts, structuring, alternative wording and conversation prep well. It handles decisions about people poorly, along with assessments that need context and anything where the cost of a mistake is high. The rule is simple: AI prepares, the manager decides.

10 tasks

1. Preparing a one-to-one

"Help me prepare a one-to-one with an employee who has missed two deadlines. Suggest 5 open questions to understand the cause, without blame."

2. Drafting feedback

"Rephrase my feedback using the SBI model: situation — behaviour — impact. Remove judgements about the person." More on the model in the guide Giving feedback to employees.

3. Goals and metrics

"Help me write a quarterly goal for the sales team in OKR format: one objective and three measurable key results."

4. A meeting agenda

"Draft an agenda for a one-hour project kick-off meeting: 4 items, time for each, who prepares what and what decision we need."

5. A difficult email

"Help me write an email to the team about postponing the launch by a month: honest, no panic, with the reasons and next steps."

6. Untangling a conflict

"Two employees disagree about priorities. Here are both positions. Help me see each person's interests and suggest three questions for a joint meeting."

7. An onboarding plan

"Draft a 30-day plan for a new project manager: what to learn, who to meet, what tasks to give each week."

8. Summarising long documents

"Condense this report into 5 key conclusions and 3 risks for a manager." Remove personal and confidential data from the text first.

9. Interview preparation

"Write 8 questions for a structured interview for a team lead role and criteria for assessing the answers."

10. A short training session

"Suggest a 30-minute session for the team on giving each other feedback: an exercise, discussion questions, a takeaway." For full programmes, try the Couchinfo AI training generator.

Three rules for safe use

1. Don't hand over confidential data. Employees' personal data, salaries and trade secrets belong only in tools approved by your company.
2. Check everything. AI makes mistakes with confidence: it invents numbers, mixes up facts and oversimplifies. Every output is a draft, not a finished decision. The Spot the AI mistakes exercise helps practise checking.
3. Decisions about people are yours. Assessment, promotion and dismissal aren't delegated to AI, even if it "advised" something.

How to bring AI into a management team's work

The hardest part isn't learning the tool but changing habits. Managers try AI once, get a mediocre result and give up. What helps is practice on real tasks: two or three hours working together, where everyone solves their own task with AI and shares their prompts.

There are two ready-made trainings for this: Performance management with AI — goals, metrics, meetings and an AI assistant for managers — and AI adoption and change — how to lead a team through the change.