AI Prompt Formulas: 5 Proven Structures That Work Every Time

System · Task · Context · Output

The gold standard for professional prompt engineering. Used by AI Prompt Architect.

System: You are a [role] with [expertise].
Task: [Specific instruction].
Context: [Background information, constraints, audience].
Output: [Format, length, tone, examples].

Best for: All-purpose — coding, writing, analysis, creative

Role · Action · Context · Expectation

Popular in marketing and content creation workflows.

Role: Act as a [professional role].
Action: [What to do].
Context: [Situation and background].
Expectation: [What good output looks like].

Best for: Marketing copy, sales emails, social media

Role · Instructions · Steps · End goal · Narrowing

Strong for complex, multi-step tasks.

Role: You are a [expert].
Instructions: [Core task].
Steps: 1. [Step 1] 2. [Step 2] 3. [Step 3].
End goal: [What success looks like].
Narrowing: [Constraints and exclusions].

Best for: Technical documentation, research, project planning

Context · Objective · Style · Tone · Audience · Response

Developed by the Singapore government for AI policy work.

Context: [Background situation].
Objective: [What to achieve].
Style: [Writing style].
Tone: [Emotional register].
Audience: [Who reads this].
Response: [Format and length].

Best for: Government, policy, formal documentation

Let's think step by step

Not a template — a reasoning technique that improves accuracy on complex problems.

"Let's approach this step by step:
1. First, identify [X]
2. Then, analyse [Y]
3. Finally, conclude [Z]
Show your reasoning at each step."

Best for: Math, logic, code debugging, complex analysis