AI for Team Leads

Leading people asks a lot of you. A short note can carry worry, praise, a change, or a promise. It can also be misunderstood.

This book helps with three small jobs: plan a 1:1, write a team update, and make a hard email clear.

You do not hand over the people part. An AI app was not in the room. It cannot know why a person acted as they did. It cannot decide who is right. It cannot make a fair staff decision for you.

It can help shape words from facts you already have.

You may be new to these apps. That is fine. Go one page at a time. Use a short copy box. Read the answer. Fix it. Keep it only if it is true and sounds like you.

For private or formal staff matters, use your workplace’s approved human process.

With care,
Everyday AI

You will use AI for one small writing job at a time.

1. Pick a page that fits today’s job.
2. Open ChatGPT, Claude, or Gemini.
3. Copy the short words in the cream box.
4. Fill in the parts inside square marks.
5. Paste the words into the app.
6. Read the answer slowly.
7. Fix or delete anything that is wrong.

AI makes a draft. A draft is a first try. It is not a decision and it is not ready just because it sounds smooth.

This book helps with 1:1 plans, team updates, hard emails, praise, meeting notes, handoffs, and corrections.

It does not score people. It does not choose discipline. It does not give employment or HR advice. It does not decide what another person meant.

Start with low-risk writing. Keep judgment, care, and final choices with you.

Prompt: The words you type or paste into an AI app.

Copy box: The cream box on a page. Copy its short lines. Replace words inside square marks with your own facts.

Paste: Place copied words into the message area. On a phone or iPad, press and hold there. Tap Paste.

Check: Read the answer and compare it with your notes. Look at every name, date, number, promise, and claim.

Official app: The real app or website your workplace allows you to use. Ask if you are not sure.

Never paste passwords. Do not paste medical details, pay files, private reviews, complaints, or secret business information. Follow your workplace’s rules.

If a message could affect someone’s job, pay, leave, safety, or formal record, stop. Use the approved human process.

What you are doing: You are giving the app facts, not guesses.

Ravi’s note says, “Draft came Tuesday. Maybe she does not care.” The first part can be checked. The second part is a guess.

1. Open your notes.
2. Underline what you saw, heard, received, or agreed.
3. Circle dates and promises.
4. Mark guesses with a question mark.
5. Remove private details the app does not need.
6. Paste only the useful facts.

Do not: Treat the app’s wording as proof of a person’s motive.

Next: Compare both lists with your notes. Delete anything that did not come from you.

Stop here if your notes involve a complaint, health issue, formal review, pay, leave, or possible discipline.

What you are doing: You are making a short plan so you can listen well.

Ravi has a 1:1 with Mei. The launch slipped by one day. Mei asked for clearer owners. Ravi promised to check with design.

1. Find the notes from the last talk.
2. Write down any promise you made.
3. Add one topic the other person raised.
4. Add one topic you need to discuss.
5. Leave space for a question.
6. Keep the plan to five short lines.

Do not: Let a tidy plan turn into a hidden performance review.

Next: Circle your promise. In the meeting, ask and listen. Do not read the draft like a script.

Stop here when the plan is easy to scan.

What you are doing: You are telling the team what is done, open, and needed.

Keisha has three facts. The search fix shipped. Two interviews happened. The handoff note is still open.

1. Make three headings: Done, Open, Need.
2. Put each fact under one heading.
3. Name an owner only when confirmed.
4. Use a date only when agreed.
5. Remove praise words that hide the real state.
6. Keep the note short.

Do not: Allow cheerful wording to cover an open problem.

Next: Check each bullet against a note, ticket, calendar item, or direct talk.

Stop here when every line can be proved.

What you are doing: You are putting the change first and the action last.

Tomas must tell the group that the meeting moved to Thursday. His first draft hides the change under six lines of background.

1. Write the change in one plain sentence.
2. Add only a reason you know is true.
3. Say what has not changed.
4. Name the action readers need to take.
5. Check the date and time yourself.
6. Read the first and last lines together.

Do not: Let the answer invent a cause, promise, or rule.

Next: Read only the first and last sentences. They should show the change and action.

Stop here if the email concerns formal conduct, employment, safety, or a protected complaint.

What you are doing: You are thanking someone for a real moment you observed.

Mei saw a teammate fix an export problem. The launch file then went out. She wants to say thank you without hinting at a reward.

1. Name what you saw.
2. State the result you can support.
3. Say what you appreciated.
4. Choose a private note or team message.
5. Remove labels about personality.
6. Keep reward promises out.

Do not: Add a rank, score, promotion hint, made-up quote, or claim about character.

Next: Ask, “Did I see this, or did the app infer it?” Keep only what you know.

Stop here when the note names the moment and says thank you.

What you are doing: You are preparing fair questions without choosing a winner.

Two people disagree about who owned a document. Each remembers the handoff in a different way.

1. List facts both people agree happened.
2. List each person’s account on its own.
3. Mark points that are still disputed.
4. Write the result the team needs now.
5. Prepare one question for each person.
6. Plan how you will record the next step.

Do not ask the app who is telling the truth. It cannot investigate. It may make one account sound settled.

A useful opening is: “I want to understand what happened and agree on the next step. I will start with what we know. Then I want to hear each view.”

Do not: Use guessed motives, labels, or invented quotes.

Next: In the room, listen and ask. Use your workplace’s process for conduct, safety, harassment, or formal action.

Stop here before your prep starts sounding like a verdict.

What you are doing: You are recording a decision that has already been made.

A launch will happen Thursday. A new color will wait. Two people missed the talk.

1. Write the decision in one sentence.
2. Say what is not included.
3. Name who made the decision accurately.
4. Add an owner only if one was named.
5. Add the confirmed date.
6. Say where questions should go.

Do not: Suggest that everyone agreed if no such agreement happened.

Next: For a costly or risky choice, ask someone who was present to check the facts.

Stop here when the note records the choice without reopening it.

What you are doing: You are showing the current state and asking for one choice.

Imani’s team has more work than it can finish by the launch date. Her manager must choose between cutting the scope and moving the date.

1. List the current facts.
2. Explain the risk without making it certain.
3. Remove guesses about effort or skill.
4. Name the choice your manager can make.
5. State when you need the choice.
6. Keep the message short.

Do not: Make the case stronger with made-up numbers or blame.

Next: Highlight your request. Make sure it asks for a choice your manager can make.

Stop here when the facts, risk, and choice are easy to find.

What you are doing: You are replacing a status meeting with a short written update.

The search fix shipped. The owner is known. No one needs to discuss or decide anything live.

1. Read the purpose on the invite.
2. Ask whether a live choice is needed.
3. Ask whether people must solve a problem together.
4. If both answers are no, write the update.
5. Name one open item, if needed.
6. Cancel the invite after checking.

Do not: Let the app create an agenda just because meetings sound careful.

Next: Check local practice. Then send the update and change the calendar.

Stop here if a live choice or hard talk still needs the room.

What you are doing: You are describing a gap in the work without judging the person.

A brief was due Monday with a risk section. It arrived Wednesday without that section. “Careless” is a thought, not an observed fact.

1. Write the agreed expectation.
2. Write what you received.
3. Add the impact only if supported.
4. Remove words about character or motive.
5. Prepare two questions.
6. State what is needed next time.

A useful opening is: “We agreed on Monday with a risk section. I received it Wednesday without that section. I want to understand what got in the way and agree on the next check-in.”

Do not: Ask an app to score the person, diagnose behavior, threaten action, or choose a consequence.

Next: Use words about the work and the agreement. Listen before deciding what comes next.

Stop here if this may become a formal review or conduct matter.

What you are doing: You are correcting your own error quickly and clearly.

Claire sent the wrong date. She fixed the calendar. The team needs the right date, not a long speech.

1. State that you were wrong.
2. Give the correct fact.
3. Say you are sorry.
4. Name the fix you made.
5. Tell readers if they need to act.
6. Correct it where the error appeared.

Do not: Shift blame, invent an excuse, overstate harm, or ask for comfort.

Next: Check that “I” owns the mistake. Check the corrected fact once more.

Stop here when the note owns, corrects, and fixes the error.

What you are doing: You are giving the covering lead only what they need.

Ravi will be away for three days. Mei will cover the launch. His notes include dates, owners, worries, and a guess about someone’s mood.

1. List the cover dates.
2. List active work and confirmed owners.
3. Add real deadlines.
4. State which choices Mei may make.
5. State what must wait.
6. Delete private comments that do not help.

Do not: Include health details, private feedback, mood guesses, complaints, or unapproved authority.

Next: Ask, “Does Mei need this to do the work?” Delete details she does not need.

Stop here when the handoff is useful without becoming gossip.

What you are doing: You are fixing one clear problem or starting again.

A smooth answer can still be false. Name the problem in one short line.

• “Use only facts from my notes. Delete all guesses.”
• “Mark anything you added. Then remove it.”
• “Use short words and short sentences.”
• “Put confirmed facts apart from open questions.”
• “This decision is made. Do not add choices.”
• “Remove promises I did not give you.”
• “Describe the work. Do not judge the person.”

Compare the new answer with your notes. A second answer can also be wrong.

If the app keeps inventing facts, close the chat. Start again with fewer notes. You can also write the message yourself.

Do not: Keep polishing a bad answer until you forget what your notes said.

Next: Save only true lines you would say face to face.

Stop here if the message affects a job, pay, leave, safety, formal performance, or conduct.

What you are doing: You are taking back full control of the message.

1. Facts: Can you point to a note, ticket, calendar item, decision, or talk behind each claim?
2. People: Did the draft add a motive, trait, score, rank, or diagnosis?
3. Promises: Did it create a promise about time, money, staff, rewards, or results?
4. Privacy: Does this reader need every detail?
5. Voice: Would you say these words in the room?
6. Action: Is the next step clear and agreed?
7. Names and dates: Did you check each one?

If one answer is wrong, edit the draft or do not send it.

The app did not earn your team’s trust. You did. A clean paragraph cannot replace listening, care, or a fair process.

Do not: Send because the answer looks polished.

Next: Put it down for one minute. Read it again as the person receiving it.

Stop here when it is true, needed, kind, and clear.

These sources support a careful writing habit. They do not show that AI can judge staff or replace a team lead.

OpenAI, “How People Are Using ChatGPT,” September 15, 2025. Writing is a large part of work use. Many writing requests ask the app to change text supplied by the user.

https://openai.com/index/how-people-are-using-chatgpt/

Anthropic Docs, “Long Context Prompting Tips.” Anthropic recommends placing long material before the request for some long-document work. Its tests found that order could improve results.

https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/long-context-tips

Anthropic Academy, “The AI Fluency Index.” Repeated checking and refining are linked with stronger AI-use habits. This is a link, not proof of cause.

https://academy.claude.com/tutorials/the-ai-fluency-index

The lesson is simple: begin with your facts, ask for a small edit, and check the answer. Keep people judgments and formal choices with humans.