AI assistants can draft emails, summarize documents, explain code and brainstorm ideas, but the quality of the answer depends heavily on the quality of the request. A vague prompt gets a vague answer. This guide gives you a simple five-part framework, before-and-after examples and a few habits for checking what the assistant gives back.
The advice applies to most general-purpose assistants, including ChatGPT, Claude and Gemini, because they all respond to the same basics: clear context, a specific task and a defined output.
The five-part prompt framework
You do not need every part every time, but each one removes guesswork.
- Role or perspective. Tell the assistant who it should act as, for example an experienced copy editor or a patient tutor for a beginner.
- Context. Explain the situation, audience and background. Assistants do not know your business, your reader or what you have already tried.
- Task. State the exact action: summarize, compare, rewrite, brainstorm ten options, explain step by step.
- Format. Say how you want the result: a table, five bullet points, an email under 120 words, a checklist.
- Constraints. Add limits such as tone, length, words to avoid, sources to rely on or what not to change.
Before and after examples
| Weak prompt | Stronger prompt |
|---|---|
| Write a product description. | You are an e-commerce copywriter. Write a 60-word product description for a stainless steel water bottle aimed at hikers. Friendly tone, no exaggerated claims, end with one call to action. |
| Explain investing. | Explain index funds to a complete beginner in under 200 words. Use one everyday analogy and finish with three questions I should ask before investing. |
| Fix my email. | Rewrite the email below to be shorter and more polite. Keep the deadline of Friday and keep my request for a reply by Wednesday. |
Notice the pattern: the stronger versions state audience, length and what must stay unchanged.
Give examples of what good looks like
If you want a particular style, show it. Paste a short sample of writing you like or a template you use and ask the assistant to follow that structure. Providing one or two examples is often more effective than trying to describe the style in words.
Ask for a plan before a long answer
For complex work, ask the assistant to outline first. You can correct the direction cheaply, then ask it to expand each section. This avoids receiving two thousand words that miss the point.
Iterate instead of starting over
Treat the first answer as a draft. Useful follow-ups include:
- Make it shorter and more direct.
- Give me three alternatives with different tones.
- What assumptions did you make? List them.
- Explain that as if I were new to the topic.
- Point out weaknesses or missing information in your own answer.
Check the output
AI assistants can produce confident statements that are wrong, outdated or invented, including made-up statistics and sources. Build simple checks into your routine:
- Verify facts, numbers, quotes and links against trustworthy sources.
- Ask the assistant which parts it is least sure about.
- Be extra careful with medical, legal and financial topics, where mistakes are costly.
- Run any code or formulas before relying on them.
Protect your information
Do not paste passwords, confidential client data, personal identification details or anything you are not allowed to share. Check the privacy settings and data policy of the tool you use, especially for work. A good habit is to anonymize names and figures before you paste text.
Prompts to try today
- Summarize this article in five bullet points and list two questions it leaves unanswered.
- Act as an interviewer and ask me five questions to prepare for a job interview in my field.
- Turn these meeting notes into a task list with owners and deadlines.
- Compare option A and option B in a table covering cost, effort and risk.
Final thoughts
Better prompts come from clearer thinking. If you cannot explain what a good answer looks like, the assistant cannot either. Start with context, a precise task and a format, then refine. To see where these tools are heading, read our overview of AI trends worth watching, and browse the Productivity section for more ways to save time.