How to Write Better AI Prompts: 8 Techniques That Actually Work

The single fastest way to get better results from any AI model is to write a better prompt. The same model can produce a vague, generic answer or a sharp, useful one depending entirely on how you ask. These eight techniques work across ChatGPT, Claude, Gemini and every other major model.

1. Give the model a role

Start by telling the model who it should be: “You are a senior financial analyst” or “Act as a patient coding tutor.” A role sets the vocabulary, depth, and tone before you ask anything, and it consistently lifts answer quality.

2. Show an example, don’t just describe

Models learn the pattern you want far better from one or two examples than from a paragraph of instructions. If you want a specific format, paste a sample of the exact output you’re after. This is called few-shot prompting, and it’s the highest-leverage habit on this list.

3. Be specific about the output format

Say exactly what you want back: “Return a 5-row table with columns Name, Risk, Fix” beats “summarize the risks.” Specify length, structure, and style so you don’t have to re-prompt.

4. Break big asks into steps

Instead of one giant request, ask the model to work through a sequence: outline first, then draft, then refine. You can do this in one prompt (“First do X, then Y”) or across a short conversation.

5. Add context and constraints

Tell the model what it needs to know and what to avoid: the audience, the goal, the tone, the word count, words not to use. Constraints don’t limit creativity — they focus it.

6. Ask it to reason before answering

For anything involving logic, math, or analysis, adding “think step by step before giving your final answer” measurably improves accuracy, because the model works through the problem instead of guessing the conclusion.

7. Iterate — your first prompt is a draft

Treat prompting as a conversation. When the output is close but not right, don’t start over — tell the model exactly what to change: “Make it shorter, drop the jargon, keep the second example.”

8. Set the audience and tone

“Explain this to a smart 12-year-old” and “Explain this to a database engineer” produce completely different answers. Naming the reader is one of the simplest ways to control quality.

Put it together

A strong prompt usually stacks several of these: a role, an example, a clear format, and the right context. Browse our tested prompt library to see these techniques applied to real tasks you can copy today.