Context · Action · Result · Example

A clean, example-led prompt framework. Useful when you want the LLM to model a specific voice or pattern.

The components

CContext
The situation, audience, and background.
AAction
The specific action the LLM should take.
RResult
The desired outcome and how to recognise success.
EExample
A concrete example of the kind of output expected.

Worked example

Rough prompt

Write a polite but firm reply to a client who's overdue on payment.

Super-prompt produced by CARE

[Context] You are writing on behalf of a small consulting firm. The client is a long-standing relationship who has historically paid on time but is now 21 days late on a £4,500 invoice with no acknowledgement of two previous reminders.
[Action] Draft a short follow-up email that maintains the relationship but sets a clear next step. The email is from the firm's senior partner, not from accounts.
[Result] An email of 90–130 words that the client reads as warm but unmistakably serious. They should be moved to either pay within 7 days or open a conversation about the cause of the delay.
[Example] Tone reference: "I hope you're well — I'm conscious we've not yet heard back on invoice INV-0421. I wanted to write to you directly because…"

Try CARE on your own prompt.

Open the builder with CARE pre-selected and watch a rough idea become a structured super-prompt.

Build with CARE →