The RACE Framework
Role, Action, Context, Expectation
RACE is RTF with two upgrades that matter for serious work: Context and Expectation. You set the Role, name the Action, supply the Context the model can't infer, and state your Expectation for the result. The Context slot is the one most people skip — and it's usually the difference between a generic answer and one that fits your actual situation.
Last updated · By the Prompt Orange team
Best for
Business and expert tasks that need real context to come out right.
What each part means
RACE stands for Role, Action, Context, Expectation. Here's what to put in each slot.
Role
Who the AI should act as, which sets its expertise and tone.
Example: You are a B2B marketing strategist.
Action
The specific action or deliverable you want.
Example: Draft a three-email nurture sequence.
Context
The background only you know — product, audience, constraints, what's been tried. This is where most prompts fall down.
Example: Our product is a £49/month invoicing tool for freelancers; leads signed up for a free guide but haven't trialled it.
Expectation
What a good answer looks like — success criteria, tone, length, what to avoid.
Example: Each email under 120 words, friendly but not pushy, with one clear call to action.
The RACE template
Copy this, fill in the brackets, and paste it into ChatGPT, Claude, Gemini, or any AI tool.
Role: You are a [role/expertise]. Action: [The deliverable you want]. Context: [Background, audience, constraints, what's been tried]. Expectation: [What a great answer looks like — criteria, tone, length, what to avoid].
Before & after: RACE in action
See how the framework turns a vague prompt into a strong one.
“Write a nurture email sequence for our SaaS product.”
Too vague—AI has to guess what you want
“Role: You are a B2B marketing strategist. Action: Draft a three-email nurture sequence. Context: Our product is a £49/month invoicing tool for freelancers; these leads downloaded a free tax guide but haven't started a trial. Expectation: Each email under 120 words, friendly but not pushy, one clear CTA per email, and the third should create gentle urgency without fake deadlines.”
Specific, clear, ready to use
Why this works:
The Context tells the model who it's writing to, what they've done, and the price point — so the copy can reference the right pain. The Expectation sets length, tone, and an explicit 'no fake deadlines' guardrail, so you get on-brand drafts instead of pushy boilerplate.
Tips for getting the most from RACE
Treat Context as the most important slot. If the model gets something wrong, you usually left out context it needed.
In Expectation, state what to avoid as well as what you want — it prevents common failure modes.
RACE is a strong default for work prompts: marketing, analysis, planning, anything where your situation is specific.
Paste real (non-confidential) details into Context — numbers, audience, prior attempts — rather than describing them abstractly.