You do not need a complicated prompt to use AI at work. Begin with a small task you already understand, such as drafting a routine email, organising rough notes or turning a list of points into an agenda.

The aim is to get a useful first version quickly. You then check the facts, change the wording and decide whether the result is ready to use. In this lesson, you will practise that process on a simple workplace example.

Choose a task with a clear result

Start with something you can describe in one sentence. For example:

Draft an email confirming a meeting.

That tells the assistant what to produce, but it leaves important details open. Who is receiving it? What time is the meeting? What tone would suit the relationship?

Add only the information needed to complete the task:

Draft a short, friendly email to a colleague confirming our meeting on Thursday at 10 am. Ask them to send any agenda items by Wednesday afternoon.

Now you have a result you can assess. You can check the time, the request and whether the tone sounds right.

Say what you already know

AI should not fill gaps with invented details. If the location has not been agreed, do not expect it to know where the meeting will take place.

You can tell it how to handle missing information:

If any detail is missing, leave a placeholder rather than making it up.

This is especially useful for meeting notes, project updates and customer messages. A draft that clearly shows a missing detail is easier to fix than one that confidently supplies the wrong detail.

Describe the audience and format

A message to a close colleague may sound different from one to a new customer. Tell the assistant who will read the work and what form you need.

For example:

Turn these points into a short update for my manager. Use plain English, three bullet points and a final line identifying the decision needed.

You have given it an audience, a format and a purpose. Asking the AI to “act as an expert” is less useful than giving it those concrete details.

Provide suitable material

AI can help organise information you supply, but think before pasting workplace content into a tool. Your employer may have rules about customer data, confidential documents or approved AI services.

For practice, use fictional details. At work, use a tool approved for the information involved and remove names or sensitive details when they are unnecessary.

You can still ask for help without sharing the underlying document. For example:

Give me a simple structure for summarising a customer meeting, including decisions, actions and unanswered questions.

Review the first response

Read the answer as if a colleague had prepared it for you. Ask:

  • Are the names, dates and figures correct?
  • Has it added a claim I did not provide?
  • Does the wording fit the audience?
  • Is anything important missing?
  • Would I be comfortable sending or using this?

If the draft is close, give one clear follow-up instruction. You might say, “Make the opening more direct” or “Keep the actions, but remove the repeated background.”

Know when to stop

You can keep asking an AI assistant to revise the same text, but more versions do not always make it better. Once the work is accurate, clear and suitable for its purpose, make any final changes yourself.

That is often the quickest workflow: give a clear task, review one draft, request a focused change if needed and finish it.

Try it yourself

Use this fictional workplace task:

Draft a short email to a colleague confirming that a project update is due on Friday. Ask them to send their figures by Thursday lunchtime. Keep the tone polite and direct. Do not invent a project name.

Review the response. Check whether it includes both deadlines and whether you would send it as written. Ask for one change, then compare the two versions.

Key takeaway

A useful work prompt states the task, the relevant facts, the audience and the form of the result. It also makes clear when the AI should leave a gap rather than guess.

Choose small tasks you can judge, protect workplace information and review every result before using it. In the next lesson, we will apply those habits to emails and other everyday writing.

Next