Prompt Writing Basics: 10 Techniques That Get Better Answers From AI

If an AI chatbot has ever given you a generic, bland, or off-target answer, the problem was probably not the AI — it was the prompt. The good news: writing better prompts is a skill anyone can learn in an afternoon. Here are ten techniques that work across ChatGPT, Claude, Gemini, and other assistants.

1. Say exactly what you want

Vague in, vague out. Compare:

• Weak: “Write about email marketing.”
• Strong: “Write a 300-word introduction to email marketing for owners of small bakeries who have never sent a newsletter.”

The second prompt defines the topic, length, audience, and their starting point.

2. Give context the AI can’t guess

The model does not know your business, your customers, or what you tried last week. Tell it. A sentence or two of background — “I run a two-person web design studio and my clients are local restaurants” — dramatically improves relevance.

3. Assign a role

Starting with “You are an experienced high-school chemistry teacher” or “Act as a skeptical investor reviewing this pitch” steers tone, depth, and vocabulary. Roles are especially useful for feedback, because they give the AI permission to be critical.

4. Specify the format

Ask for exactly the shape you need: a table, a bulleted checklist, three options, a numbered step-by-step process, or an email under 120 words. If you will paste the result somewhere, say where — “formatted for a LinkedIn post” or “as a plain-text SMS”.

5. Show an example

If you have a style in mind, paste a short sample and say “match this tone and structure.” One good example often works better than a paragraph of description.

6. Break big tasks into steps

Instead of “Write my whole business plan,” go section by section: first an outline, then the market analysis, then pricing. You will catch problems earlier and get far better depth in each part.

7. Ask it to think before answering

For reasoning, math, or decisions, add: “Work through this step by step before giving your final answer.” Models are noticeably more accurate when they lay out their reasoning instead of jumping straight to a conclusion.

8. Set constraints — including what to avoid

Constraints sharpen output: “Use simple words a 12-year-old would understand,” “No jargon,” “Don’t mention pricing,” “Avoid clichés like ‘game-changer’.” Telling the AI what not to do is just as useful as telling it what to do.

9. Iterate instead of starting over

Your first result is a draft. Reply with targeted feedback: “Make the second paragraph shorter,” “More concrete examples,” or “That’s too formal — make it friendly.” The conversation keeps context, so each refinement builds on the last.

10. Ask the AI to check its own work

Finish with: “Review your answer. What’s missing, unclear, or possibly wrong?” This simple follow-up often surfaces errors and gaps. It is not a substitute for your own fact-checking — see our guide to AI hallucinations — but it catches a lot.

A reusable prompt template

Copy this structure and fill in the blanks:

Role: You are a [expert type].
Context: I am [who you are / situation].
Task: I need [specific output].
Audience: It’s for [who will read it].
Format: Give it as [table / list / email / X words].
Constraints: [tone, things to avoid, must-includes].

The bottom line

Treat the AI like a capable new colleague on their first day: smart, fast, eager — but knowing nothing about your situation. The clearer your brief, the better the work. Want to understand why these techniques work? Read how large language models actually work.