RACE model for ChatGPT prompts: the template with examples

The RACE model is one of the best-known structures for building a ChatGPT prompt: Role, Action, Context, Execute. Useful, except for two confusions. In marketing circles, RACE usually means something entirely different, and even within the prompting world three readings of that last letter circulate. This article sorts that out and then delivers what you actually came for: a ready-made template you can copy, filled in for the marketing tasks that recur most often at an SME. The theory behind good prompting is deliberately left aside; it lives in our guide to writing prompts and in the follow-up on prompt engineering.

What does the RACE model stand for in ChatGPT prompts?

RACE is a prompt template with four blocks: Role (which role the AI takes on), Action (the exact task you are asking for), Context (what the model needs to know about your situation) and Execute (the form the answer should take). You write them underneath each other in one prompt. The framework comes from the practice of AI consultants, not from academic research.

Block	Question you answer	Example line
Role	Who should the AI be?	”You are a copywriter with experience in B2B software.”
Action	What needs to happen?	”Write a follow-up email for prospects who did not reply.”
Context	What does it need to know to do it well?	”Audience, product, tone of voice, what has already been sent.”
Execute	What should the answer look like?	”Maximum 130 words, three subject lines, no bullets.”

Mind the name. Anyone talking about the RACE model in a marketing context usually means the Smart Insights planning model: Reach, Act, Convert, Engage, a breakdown of the customer journey that Dave Chaffey introduced around 2010 and that has nothing to do with AI prompts. Same four letters, different field. Here we consistently mean the prompt framework; the planning model belongs in a conversation about online marketing and channel strategy.

What does the E in RACE stand for?

For Execute, in the most widely used version: the instruction about the form and format of the answer. Two variants circulate alongside it, Expectation and Examples. All three point at the same place in the prompt, namely the closing block in which you set out what the output should look like. Variants exist because nobody ever formally fixed the framework.

Who uses what:

• Execute: the version from Trust Insights, where Christopher Penn describes the framework, and from Ayse Ozturk, who sums it up as “specify the output style”.
• Expectation: the version from Fabio Vivas, with the sentence template “As a [Role], [Action] considering [Context]. I expect [Expectation]”.
• Examples: a variant that mostly shows up in prompt libraries and uses the closing block to hand over examples instead of formatting requirements.

In practice the choice matters little, as long as you stay consistent. We stick with Execute because formatting requirements are needed in nearly every task and examples are not. Handing over examples remains useful, but as an addition, as covered further down.

What does a RACE prompt actually look like?

Like four labelled lines or blocks underneath each other, each starting with the name of the component. Those labels are not decoration: they force you to skip nothing and make the prompt reusable, because you only have to replace the fill-in fields. This is the empty template:

Role: You are a [job title or expertise] with experience in [sector or domain].
Action: [Verb] [what exactly] for [whom].
Context: We are [company or product]. Audience: [description].
         Tone of voice: [description]. What you need to know: [facts, figures,
         constraints, what has already happened].
Execute: [Length and language]. [Structure: heading, bullets, table, number of
         variants]. Avoid: [what you explicitly do not want].

And this is what it looks like filled in for a concrete task:

Role: You are a copywriter with experience in B2B software for manufacturers.
Action: Write a follow-up email to prospects who attended our webinar last month
        but have not booked a demo.
Context: We sell planning software to manufacturers with 50 to 250 employees.
         The reader is a production manager, not an IT specialist. Their biggest
         frustration is planning in spreadsheets and reshuffling every week.
         The webinar covered planning mistakes that cost money.
         Tone of voice: matter-of-fact, no superlatives, no sales talk.
Execute: Maximum 130 words, in English, with three subject lines to choose from.
         Close with one concrete question.
         Avoid: feature lists and the word "solution".

The difference with “write a follow-up email about our webinar” is not subtle. The second version leaves the model nothing to work with but an average email; the first gives it enough to write something that sounds like your company.

Which RACE templates can you copy straight away?

Below are six templates for recurring marketing tasks. Replace the brackets with your own details and leave the labels in place. They are deliberately compact; your Context block may well end up three times as long, because that is where quality comes from.

Task	RACE template
Product description	Role: copywriter for an online shop in [sector]. Action: write a product description for [product] aimed at [audience]. Context: [what the product does], [why customers hesitate], tone of voice [description]. Execute: 120 words, opening line plus three benefits plus closing line, no superlatives.
LinkedIn post	Role: content lead at [company]. Action: write three variants of a post about [topic]. Context: audience [role and sector], what we have to say about it: [position or experience]. Execute: max 120 words per variant, each with a different opening (question, figure, anecdote), no hashtags.
Follow-up email	Role: [your job title] at [company]. Action: write a follow-up email to [type of contact] who [situation]. Context: earlier contact moments [description], their objection is probably [objection], tone of voice [description]. Execute: max 130 words, first person, three subject lines, close with one question.
Ad copy	Role: performance copywriter. Action: write five ad headlines and two descriptions for [campaign]. Context: offer [description], audience [description], what sets us apart: [point]. Execute: headlines max 30 characters, descriptions max 90 characters, no exclamation marks, English.
Review analysis	Role: marketing analyst. Action: analyse the customer reviews below and pull out the recurring patterns. Context: we are a [type of business], we want to know [question]. Reviews: [paste here]. Execute: five patterns, each with a quote from the data and one action we can take, in a table.
Blog brief	Role: content strategist. Action: create an article brief about [topic]. Context: audience [description], search intent [informational or commercial], what we have already published: [links]. Execute: working title, six to eight H2s phrased as questions, two lines of direction per H2, plus five FAQ questions.

If you work with content at volume, this is exactly the point where templates turn into a process; what that looks like structurally is covered under AI content production.

How do you sharpen Role, Action, Context and Execute?

Each block has its own rule of thumb: keep the role credible, make the action measurable, fill in the context until a new colleague could get going with it, and describe the form so precisely that you can tick off the result without discussion. Most disappointing output traces back to one of these four blocks being filled in too thinly.

• Role: pick a role that matches the perspective you want (“copywriter for online shops”, not “the world’s best marketer”). Do not expect miracles from it in terms of accuracy: research by Zheng and colleagues tested 162 roles on 2,410 factual questions and found that a persona in the system prompt does not systematically improve performance. So use Role for tone and angle, and put the knowledge you need into Context yourself.
• Action: one task per prompt, with a verb up front. Ask for an email, three posts and an analysis in the same prompt and you get three half answers. Splitting is nearly always faster than reworking.
• Context: this is the block that makes the biggest difference and gets skipped most often. Audience, offer, tone of voice, what has already happened, what is off limits. A good test: could a freelancer start from this brief without calling you back?
• Execute: length, language, structure, number of variants, and explicitly what you do not want. That last part is the cheapest quality gain available; “avoid clichés such as groundbreaking and revolutionary” saves half a rewrite.

When should you add examples to your RACE prompt?

As soon as describing the form no longer conveys what you mean. Describing a style only gets you so far; two paragraphs you like say more than ten adjectives. Paste those at the bottom of the prompt, after Execute, with a short instruction alongside (“write in this voice, not about this content”).

That technique is called few-shot prompting and is exactly what the Examples variant of RACE refers to. It pays off above all for tone of voice, for a fixed format that is hard to describe (a product sheet, a reporting block) and for classification tasks, where one example per category puts the model on the right track immediately. Two to five examples usually suffice; beyond that the gain drops off and the model mostly starts copying. The broader set of techniques, from handing over examples to reasoning step by step, is covered in our article on prompt engineering. A short definition of the term is also in the glossary.

What usually goes wrong with a RACE prompt?

Rarely the framework itself, nearly always the way it is filled in. The four classics: an empty Context block, several tasks in one prompt, an Execute block that only says “professional”, and a template filled in mechanically without thinking about the task. A fifth pitfall is subtler: leaning too heavily on the role.

What helps against that is banal but effective. Treat the first answer as a rough version and steer in a second turn, rather than endlessly perfecting the prompt; a three-turn conversation usually produces better work than one mega-prompt. Check facts, figures and names before publishing, even when the answer reads smoothly. And save what works straight away, because the prompt you get right today after four attempts is gone next week. If you want to see how the choice between AI text and your own work plays out in practice, AI copywriting versus human is a useful companion piece.

How do you turn RACE into a standing way of working for your team?

By turning your best prompts into templates and storing them centrally with fill-in fields. That way everyone reaches the same quality and you build per task on what is proven to work, instead of improvising every time. One shared document is enough to start; you only need a tool once there are dozens.

Practically: pick three recurring tasks, work out one RACE template per task, test for a week, and sharpen based on what the output structurally misses. Agree on who maintains the library and add one line to each template about when to use it. Review after a month which templates actually get opened; those two or three are your real gain, the rest can go. If this grows into recurring workflows you would rather run automatically, marketing automation is the next step, and beyond that AI agents.

Want your team to master this on your own tasks? ClickForest runs practical AI training for SMEs in Flanders, in which we work out the templates together for the work that comes back every week at your company. See the AI training or book a no-obligation video call.

What should you remember about the RACE model?

RACE is not a miracle cure but a memory aid with four blocks: Role, Action, Context, Execute. The value is not in the letters but in the discipline they enforce, and above all in the Context block that most people fill in too thinly. Use the framework where it belongs, for recurring tasks you want to standardise, and do not confuse it with the identically named planning model from marketing.

If you prefer a looser formula, start with writing prompts in four steps; if you want to go beyond templates, head to prompt engineering. Which assistant you use matters less than you might think, though our comparison of AI tools helps you work out what suits which task. And if you want to use AI more broadly than text alone, that starts at AI for growth.