AI Prompting Basics: How to Get Better Answers From Any Chatbot
Most disappointing AI answers come from vague questions, not weak models. A good prompt reads like a clear brief to a capable new colleague who knows nothing about your situation. These habits work across ChatGPT, Claude, Gemini and most other assistants.
The five parts of a good prompt
- Context: who you are, who the audience is, and what the situation is.
- Task: the single thing you want done, stated as a clear instruction.
- Constraints: length, tone, things to include or avoid.
- Format: bullet list, table, email, JSON, step-by-step plan.
- Material: the text, data or notes the model should work from.
Before and after
Weak: "Write about our new product."
Better: "I run marketing for a small accounting software company. Write a 150-word announcement for existing customers about our new invoice-reminder feature. Friendly, plain English, no jargon, end with one call to action. Use the notes below."
The second prompt tells the model the audience, the length, the tone and the source material, so it has far less to guess.
Iterate instead of starting over
Treat the first answer as a draft. Reply with specific feedback such as "shorter, and lead with the benefit" or "keep section two but make the tone more formal". Targeted follow-ups usually beat rewriting the whole prompt.
Ask for reasoning on hard problems
For analysis, planning and maths, ask the model to work through the problem step by step, list its assumptions, or show how it reached the answer. It makes mistakes easier to spot. For decisions, ask for options with pros and cons rather than a single verdict.
Give examples
If you want a particular style or structure, paste one or two examples of what good looks like. Showing is more reliable than describing.
Always verify
- Check names, numbers, dates and quotes against a reliable source.
- Ask the model to say when it is unsure rather than guess.
- Do not paste confidential information unless your organisation's policy allows it.
Up next
AI Glossary: LLMs, Tokens, Agents, RAG and Hallucinations Explained Simply
Plain-English definitions of the AI terms you keep seeing in the news: LLM, token, context window, hallucination, RAG, fine-tuning, agents and more.
Read next guide →Keep reading
All guides →What Are AI Agents? A Beginner's Guide to How They Work and Where They Fail
AI agents plan steps and use tools to finish tasks for you. Learn how they work, where they help, where they break, and how to try one safely.
Read the guide →How to Spot AI-Generated Fakes: A Practical Checklist for Images, Video, Audio and Text
AI fakes are getting harder to detect by eye. Use this checklist of visual clues, source checks and verification habits before you share anything.
Read the guide →AI for Small Business: 7 Practical Ways to Save Time Without Big Risks
Seven realistic ways small teams use AI tools today for writing, customer service, research, bookkeeping prep and more, plus how to start safely.
Read the guide →