What this is: A repeatable structure you can add to any prompt inside Claude. It turns vague requests into specific, high-quality outputs. Use this any time you want to improve your prompts
5 minutes to learn. Then you’ll use it every day.
Claude is a large language model. The more context you give it, the better the output. This 5-step structure works for every type of prompt, you don't need to use the exact prompt. It's just a guide that will help you get the best out of Claude if you choose to use it in your prompts.
The most important thing you can do with a prompt is to use your own language and ask for exactly what you want from Claude to get the most out of it.
You are a top 0.1% expert in [X field].
This tells Claude what lens to think through. Be specific. This sets Claude up as an expert for you in whatever topic you're requesting Claude to help you with.
Example: You are a top 0.1% expert in content strategy for online coaches.
Here’s the background: [who you are, what you’re building, why it matters, who it’s for].
This is where you drop in the real detail. Claude has no memory between chats unless you give it context.
Example: I run an AI education business for female entrepreneurs. I’m building a 5-day email sequence to launch my new workshop. My audience are coaches and creatives who are scared of tech.
Return your response as [bullet points / a table / a script / step-by-step instructions / a sales email].
Tell Claude exactly how to package the output. If you don’t, it decides for you.