Articles Learn AI About
Level 2 — Starting to use
Page 1 of 5 · Starting to use

How to write a good prompt

6 min read · No prior knowledge needed
After this page — You'll be able to consistently get high-quality, specific output from any AI tool.

Why your output is only as good as your input

Most people who feel disappointed by AI are disappointed by their own prompts, not by the AI. A vague question gets a vague answer. A specific, well-structured prompt gets a specific, useful response.

Prompting is a learnable skill. It takes maybe an hour of practice to get noticeably better at it. Here’s what actually matters.

The four elements of a strong prompt

1. Role — tell AI who it is

Starting a prompt with a role dramatically changes the output. “You are an experienced HR manager” produces different answers than “you are a recent graduate.” The role sets the frame for everything that follows.

Compare these

BAD: "How do I ask for a raise?"

GOOD: "You are an experienced career coach. I've been in my current role for 18 months and have taken on significantly more responsibility. How should I approach asking for a salary increase? My manager tends to focus on budgets."

2. Context — give it the information it needs

AI only knows what you tell it. The more relevant context you provide, the more tailored the response. This is the most common thing beginners skip.

3. Task — be specific about what you want

“Help me with my email” is too vague. “Rewrite this email to be more professional and reduce it to under 100 words” is specific enough to produce useful output.

4. Format — tell it how to respond

Do you want bullet points or prose? A table? A numbered list? Short paragraphs? If you don’t specify, AI will choose — and it often chooses something that doesn’t suit your purpose.

The before and after test

Before"Write me a LinkedIn post about my new job."
After"Write a LinkedIn post about starting a new role as a product manager at a fintech startup. Tone: genuine and excited, not corporate. Mention that I'm looking forward to meeting people in the industry. Under 100 words. No buzzwords."

The “after” prompt will produce something you might actually post. The “before” prompt will produce something you’ll immediately delete.

The most powerful technique: show it an example

If you have an example of the kind of output you want — a previous piece of writing in your style, a format you like — paste it in. Tell AI: “Write something in this style” or “format your response like this example.”

This works better than any amount of description. AI learns from examples far more reliably than from instructions.

When the output isn’t right: iterate

A single prompt rarely produces perfect output. The skill isn’t writing one perfect prompt — it’s knowing how to refine.

Try these follow-up instructions:

  • “Make it shorter”
  • “Make it less formal”
  • “Add a concrete example”
  • “Rewrite the opening — it sounds too generic”
  • “Try again but this time focus on [specific aspect]”

Each iteration gets you closer. Think of it as a collaboration, not a one-shot transaction.

▶ Practice — takes 5 minutes

Take something you actually need to write today — an email, a message, anything. Write your first instinct prompt, see the result, then rewrite the prompt using the four elements above. Compare the two outputs. The difference will be obvious.

Next in Starting to use

Prompt templates for real situations →