Role prompting: how to tell an AI who it is

The single fastest way to raise output quality: tell the model who it is before what to do.

Why it works

A role primes the model's tone, vocabulary, and standards. 'You are a senior security engineer' pulls in caution, threat-modelling, and precision that 'help me with code' never will. Production agents (Cursor, Devin, Claude) all open by establishing an explicit persona.

How to do it

Open with: 'You are a <seniority> <domain> expert.' Be specific about seniority and domain: 'senior B2B copywriter' beats 'writer'. Add a one-line standard if it matters: '...who writes concise, jargon-free copy.'

Try it

Rewrite this so it sets a clear, specific role. Aim for a score of 70+.

57structure

Target: 70. Keep working the prompt. The score moves as you type.

Role / persona: Open with a persona, e.g. "You are a senior backend engineer."
Constraints: State limits and do/don't rules, e.g. "Keep it under 200 words. Do not invent APIs."
Output format: Specify the exact shape, e.g. "Return JSON: { title, steps[] }" or "Use a markdown table.".

Now put it to work: score a prompt you actually use on the desk and see whether this part is one it was missing. This is one of seven parts. To see all of them already in place, the templates are the finished shape; to practise on something new, the daily drill posts one brief a day.