Step by Step
R
Role — who the model should be
Establishing a specific role/persona, such as "you are an expert data scientist with 10 years of experience," calibrating the model's expertise level and perspective.
Example: "You are an expert data scientist with 10 years of experience" as the role component of a prompt.
C
Context — the relevant background
Providing the relevant background information the model needs, such as "I am building a fraud detection system for a bank."
Example: "I am building a fraud detection system for a bank" as the context component, establishing the relevant background situation.
T
Task — the specific action requested
Clearly stating the specific task being requested, such as "analyze this transaction data and identify anomalies."
Example: "Analyze this transaction data and identify anomalies" as the clear, specific task component of the prompt.
F
Format and Constraints — how the answer should look and what limits apply
Format specifies the desired response structure ("respond with a bulleted list, then a confidence score"); Constraints specify limits ("use only Python, keep explanation under 200 words"). Not every prompt needs all five elements, but missing context is the single most common cause of prompt underperformance.
Example: "Respond with a bulleted list, then a confidence score. Use only Python. Keep explanation under 200 words." combining format and constraints together in one prompt.
Applied Walkthrough
1
A prompt asking simply "analyze this data" produces a vague, unfocused response that doesn't meet the user's actual needs.
2
Ask: which of the five elements (Role, Context, Task, Format, Constraints) is most likely missing here, and what's the most common single culprit? Missing context is the most common cause of prompt underperformance — the model doesn't know the broader situation or purpose behind the request.
3
Rewriting the prompt to include role ("You are an expert data scientist"), context ("I am building a fraud detection system for a bank"), a clear task, desired format, and any relevant constraints produces a dramatically more useful, targeted response.
4
This five-element framework — RCTFC — provides a systematic way to diagnose exactly why an underperforming prompt might be falling short, rather than guessing randomly at what to change.
Exam Application
Exams test whether you can name and describe all five elements of a high-quality prompt (Role, Context, Task, Format, Constraints) and whether you know that missing context is specifically identified as the single most common cause of prompt underperformance.
⚠ Common Trap
The most common trap is assuming every prompt needs all five elements explicitly stated. Not every prompt requires all five — but knowing this five-element framework specifically helps diagnose WHY an underperforming prompt might be falling short, with missing context being the single most common specific culprit.
✓ Quick Self-Check
1. What does the Role element of a prompt establish?
A specific persona or expertise level for the model, such as "you are an expert data scientist."
Tap to reveal / hide
2. What does the Context element of a prompt provide?
The relevant background information the model needs to understand the situation.
Tap to reveal / hide
3. What does the Task element of a prompt specify?
The specific action being requested.
Tap to reveal / hide
4. What do the Format and Constraints elements specify, respectively?
Format specifies the desired response structure; Constraints specify limits like language or length.
Tap to reveal / hide
5. What is identified as the single most common cause of prompt underperformance?
Missing context.
Tap to reveal / hide