The RTCF Framework
A reliable structure: Role, Task, Context, Format.
- ✓The four parts of a strong prompt
- ✓How each part changes the output
- ✓A reusable template you can apply anywhere
Explanation
A dependable way to structure prompts is RTCF.
Role — tell the model who to act as ('a senior data analyst'). This sets vocabulary and depth.
Task — state exactly what to do, with a verb ('summarise', 'compare', 'draft').
Context — give the background it needs: audience, constraints, source material.
Format — describe the output shape: bullets, a table, JSON, word count. Specifying format is the single biggest quality boost most people miss.
Code Example
| 1 | |
| 2 | |
| 3 | |
| 4 | |
Teams save RTCF prompts as reusable templates so anyone gets consistent, on-brand output.
- • Skipping Format and then being surprised the answer is a wall of text.
Write an RTCF prompt that turns a messy meeting note into a clean action-item list.
1. Which RTCF element most improves output structure?
2. The 'Role' part of a prompt mainly sets...