Reusable Prompt Structures

Prompt patterns are proven structures that consistently produce good results. Instead of reinventing prompts each time, you can adapt these patterns to your specific needs. Think of them as recipes—adjust ingredients, but the basic method works.

The Persona Pattern:

Assign the AI a specific role or expertise level to frame its responses.

Structure:

You are a [role] with expertise in [domain].
Your task is to [action] for [audience].
[Specific instructions]

Examples:

Teaching example:
'You are a patient math tutor working with a high school student struggling with algebra. Explain how to solve quadratic equations using simple language and real-world analogies. Check for understanding after each step.'

Business example:
'You are a senior business strategist with 20 years of experience in retail. Analyze this company's market position and suggest 3 strategic moves for the next year. Consider current e-commerce trends and consumer behavior shifts.'

Creative example:
'You are a creative director at a top advertising agency. Brainstorm 10 creative campaign concepts for a sustainable fashion brand targeting Gen Z consumers. Each concept should include a tagline and brief execution idea.'

Why it works: The AI's training includes text written by people in various roles. Activating a specific persona helps the model generate responses consistent with that expertise and perspective.

The Instructional Pattern:

Step-by-step explanation requests, perfect for learning and understanding.

Structure:

Explain [concept] as if I am [audience level].
Break it down into [number] simple steps.
Use [analogy/example type] to illustrate.
Examples:
Technical learning:
'Explain how blockchain technology works as if I am a small business owner with no technical background. Break it down into 5 simple steps. Use everyday analogies (like banking, record-keeping, or document signing). Avoid terms like 'cryptographic hash' or 'Merkle trees' without explaining them first.'
Skill development:
'Explain the fundamentals of SEO as if I'm launching my first website. Break it into 4 main categories. Use concrete examples from a fictional bakery website. Avoid advanced tactics—focus on what I can implement in my first month.'
Why it works: Explicitly setting knowledge level prevents both over-simplification and overwhelming complexity. The AI calibrates explanations appropriately.
The Comparison Pattern:
Structured analysis of multiple options to support decision-making.
Structure:
Compare [Option A] and [Option B] for [specific use case].
Consider these factors: [factor 1], [factor 2], [factor 3].
Format as [table/list/pros-cons].
Provide a recommendation for [specific scenario].
Examples:
Tool selection:
'Compare Notion and Asana for managing a content marketing team of 5 people. Consider: ease of use, collaboration features, content calendar capabilities, and pricing. Format as a table. Provide a recommendation for a team that creates blog posts, social media content, and newsletters.'
Strategy evaluation:
'Compare organic social media growth versus paid advertising for a new DTC skincare brand with a $50K marketing budget. Consider: timeline to results, cost per acquisition, sustainability, and brand building. Provide pros and cons for each. Recommend a split between organic and paid for the first 6 months.'
Why it works: Structured comparison prevents AI from giving generic advice. Specific evaluation criteria produce actionable insights.
The Template Fill Pattern:
Provide a structure with placeholders for the AI to complete.
Structure:
Fill in this template with [specific content]:

[Template with blanks/sections]

Guidelines: [formatting rules, tone, length]
Examples:
Social media template:
'Fill in this LinkedIn post template for announcing our new product feature:

🎉 We're excited to announce [feature name]!

[One-sentence description of what it does]

 Why we built this:
[2-3 sentences about customer pain point]

What makes it special:
• [Benefit 1]
• [Benefit 2]
• [Benefit 3]

[Call to action]

Product: Project management tool with AI-powered task prioritization
Feature: Smart deadline suggestions based on team capacity
Tone: Professional but enthusiastic
Length: Keep it skimmable—under 150 words'
Email template:
'Fill in this cold outreach email template:

Subject: [Benefit-focused subject line under 50 characters]

Hi [First Name],

[Opening line that shows you researched them]

[One sentence about the problem you notice they might have]

[One sentence about how you solve it]

[Brief social proof or credibility element]

[Soft call to action]

Best,
[Name]

Context: Selling data analytics consulting to e-commerce companies with $5M+ revenue
Tone: Professional, helpful, not salesy
Keep total email under 100 words'
Why it works: Templates give AI a clear structure to follow, reducing generic outputs and ensuring consistent formatting.
The Chain of Thought Pattern:
Ask AI to show its reasoning process before providing a final answer.
Structure:
Solve [problem].
First, break down the problem into steps.
Then, work through each step showing your reasoning.
Finally, provide your conclusion.
[Optional: Check your work for errors]
Examples:
Problem-solving:
'Our blog traffic dropped 40% last month. First, list the possible causes. Then, for each cause, explain how we could investigate whether it's the real issue. Finally, recommend the top 3 things to check first and why.'
Mathematical/logical:
'Calculate the ROI of hiring a content writer at $60K/year who produces 2 blog posts per week. First, estimate the traffic and conversion value per post based on industry averages. Show your calculations step by step. Then, determine how many months until the hire pays for itself. Finally, recommend whether this is a good investment.'
Why it works: Forcing the AI to explain reasoning produces more accurate answers. It's less likely to make logical leaps or errors when showing its work.
The Constraint Satisfaction Pattern:
Define multiple requirements that must all be met.
Structure:
Create [output] that meets ALL of these requirements:
1. [Requirement 1]
2. [Requirement 2]
3. [Requirement 3]
4. [Requirement 4]
[Add verification: 'Confirm each requirement is met.']
Examples:
Content creation:
'Write an Instagram caption that meets ALL these requirements:
1. Exactly 2,200 characters (at the character limit)
2. Includes a hook in the first sentence that stops scrolling
3. Contains 3 actionable tips
4. Ends with a question to drive engagement
5. Incorporates 5-7 relevant hashtags naturally in the text
6. Uses line breaks for readability
7. Includes one emoji per section (but not excessive)
8. Written for aspiring photographers learning mobile photography

Topic: How to take better photos with your smartphone

After completing, verify each requirement is met.'
Why it works: Explicitly listing requirements and asking for verification ensures nothing is overlooked. The AI treats each as a hard constraint.
The Refinement Pattern:
Iteratively improve outputs through multiple rounds.
Structure:
Round 1: Create [initial output] with [basic requirements]
Round 2: Now improve it by [specific enhancement]
Round 3: Make it even better by [another enhancement]
[Continue as needed]
Examples:
Writing improvement:
'Round 1: Write a product description for wireless earbuds. Focus on key features: battery life, sound quality, comfort. Keep it to 3-4 sentences.

[After receiving output]

Round 2: Now rewrite it to be more emotionally compelling. Instead of listing features, describe the experience and feelings. Use sensory language.

[After receiving improved output]

Round 3: Make it punchier. Cut any unnecessary words. Each sentence should have impact. Aim for grade 8 reading level for accessibility.'
Why it works: Complex outputs are easier to develop iteratively than to demand perfection in one prompt. Refinement allows progressive improvement.
The Perspective Shift Pattern:
Ask for the same content from different viewpoints.
Structure:
Analyze [topic] from the perspective of:
1. [Stakeholder 1]
2. [Stakeholder 2]
3. [Stakeholder 3]

For each perspective, explain:
- Their primary concern
- What they care most about
- What they're afraid of
- What would make them happy
Examples:
Product development:
'Analyze the decision to add a subscription pricing tier to our software from three perspectives:

1. Existing customer (currently paying one-time $99)
2. Company CFO (focused on recurring revenue)
3. Support team (handling customer questions)

For each, explain what concerns they'd raise and what would make them support this change.'
Why it works: Multi-perspective analysis reveals blind spots and helps you anticipate objections or opportunities you might miss from a single viewpoint.
The Socratic Pattern:
Instead of giving answers, AI asks questions to guide your thinking.
Structure:
I need to [make a decision/solve a problem].
Instead of telling me what to do, ask me 5-7 thought-provoking questions that will help me think through this clearly.
After I answer, provide 3-5 more follow-up questions based on my responses.
Finally, help me synthesize my answers into a decision.
Examples:
Career decision:
'I'm trying to decide whether to take a new job offer. Instead of advising me directly, ask me questions that will help me think through this decision clearly. Start with 5 questions about my priorities, current situation, and concerns. After I answer, ask follow-up questions based on my responses. Finally, help me synthesize my answers into clarity about what matters most.'
Why it works: For personal or complex decisions, questions are more valuable than answers. This pattern helps you clarify your own thinking rather than outsourcing the decision.
The Output Transformation Pattern:
Take existing content and transform it into different formats.
Structure:
Transform this [original format] into [new format]:

[Original content]

Adaptations needed:
- [Change 1]
- [Change 2]
- [Change 3]
Examples:
Content repurposing:
'Transform this 2,000-word blog post into a LinkedIn carousel (slide-by-slide breakdown):

[Paste blog post]

Create 8-10 slides:
- Slide 1: Hook that makes people want to swipe
- Slides 2-8: One key point per slide, 20-30 words each
- Last slide: Call-to-action

Make each slide work standalone but flow together. Keep language concise and visual.'
Format adaptation:
'Transform this technical documentation into a beginner-friendly video script:

[Paste documentation]

Adaptations:
- Conversational tone (write like you're speaking)
- Add analogies for technical concepts
- Include clear section breaks for video chapters
- Suggest where to show screen examples
- Aim for 5-minute video (roughly 750 words)'
Why it works: Explicit transformation instructions prevent the AI from just summarizing. You get genuine adaptation to new medium and audience.
Building Your Prompt Library:
Create personal templates for your common tasks:
Audit your workflows: What content do you create repeatedly?
Identify patterns: Which prompts worked well in the past?
Generalize successful prompts: Replace specific details with [placeholders]
Document in a centralized location: Notion, Google Docs, or dedicated prompt management tools
Test and refine: Try your templates in different scenarios
Version your templates: Keep track of improvements over time
Template organization structure:
📁 Prompt Library
  ├── 📄 Marketing
  │   ├── Email campaigns
  │   ├── Social media posts
  │   └── Ad copy
  ├── 📄 Content Creation
  │   ├── Blog posts
  │   ├── Video scripts
  │   └── Newsletters
  ├── 📄 Business
  │   ├── Meeting summaries
  │   ├── Proposals
  │   └── Reports
  └── 📄 Learning
      ├── Concept explanations
      ├── Skill practice
      └── Study guides
Treat your prompt library as a living document that grows with your experience. The best prompt engineers have dozens of refined templates for recurring tasks.