AI PRD Workflow

A collection of reusable AI prompts to streamline your product development workflow. Copy and paste these prompts into any AI assistant (Claude, ChatGPT, Cursor, Windsurf, or any AI-powered IDE) to automate common tasks in your product development process.

Quick Start

1. Pick a prompt from Available Prompts below
2. Copy its contents (or use ./copy-prompt.sh --list to browse)
3. Paste into any AI assistant with your project context attached

New here? Check the examples/url-shortener folder for sample outputs from every step.

[!TIP] Why PRD & RFC Driven Development? This approach significantly improves AI collaboration by providing clear boundaries, reducing hallucinations, and ensuring consistent understanding between humans and AI. Jump to full benefits

Table of Contents

• Benefits of PRD & RFC Driven Development
• Available Prompts
• Recommended Workflow
• Examples
• How to Use
• Compatibility
• Quick Tips
• Contributing
• License

Benefits of PRD & RFC Driven Development

Following a structured PRD & RFC driven development approach offers significant advantages, especially when collaborating with AI tools:

Enhanced AI Collaboration

• Clearer Instructions: Well-defined PRDs and RFCs provide AI with precise context, resulting in more accurate implementations
• Reduced Hallucinations: Structured documentation minimizes AI's tendency to fill in gaps with incorrect assumptions
• Consistent Mental Model: Both humans and AI work from the same documented understanding

Improved Development Process

• Boundary Setting: Clearly defined scope prevents AI from implementing unwanted features
• Incremental Verification: Breaking work into RFCs allows validation at each step
• Traceability: Each implementation traces back to specific requirements
• Reduced Rework: Clear specifications minimize major revisions

Business Benefits

• Predictable Outcomes: Structured approach leads to more predictable timelines and results
• Knowledge Preservation: Documentation persists even as team members or AI tools change
• Stakeholder Alignment: PRDs create shared understanding between business, technical teams, and AI
• Faster Onboarding: New team members or AI tools quickly understand goals and constraints

AI-Specific Advantages

• Context Window Optimization: Focused RFCs help AI work within context window limitations
• Targeted Expertise: Different aspects can be directed to specialized AI models or prompts
• Quality Control: Structured documentation provides clear criteria for AI to self-evaluate

Available Prompts

Recommended Workflow

Initial Development

1. Create PRD - Start with a vague idea and develop it into a complete PRD using the Interactive Creation prompt
2. Verify PRD - Identify critical gaps and improve quality using the Comprehensive Verification prompt
3. Extract Features - Transform your verified PRD into organized features with priorities and acceptance criteria
4. Create Rules - Establish technical guidelines and standards based on your PRD and features
5. Generate RFCs - Break down the project into logical, manageable implementation units
6. Implement RFCs - Use the implementation template for each RFC to guide development
7. Code Review - Review each implementation against its RFC specification and project standards
8. Testing Strategy - Generate and execute a comprehensive test plan for implemented features

Managing Changes

When new requirements or changes arise during development:

1. Analyze Changes - Use the Change Management prompt to assess impact and integration strategy
2. Update Documents - Revise affected PRD, features, rules, and RFCs based on the analysis
3. Continue Implementation - Resume development with the updated documentation

Examples

The examples/ folder contains complete sample outputs for a URL shortener project, demonstrating each step of the workflow:

• PRD - Product Requirements Document
• Features - Extracted features with MoSCoW prioritization
• Rules - Development standards and guidelines
• RFCs - Implementation units (3 RFCs)

Use these as a reference for what each prompt produces.

How to Use

Method 1: Manual Copy

1. Open the desired prompt file
2. Copy the entire contents
3. Paste into your AI assistant
4. Attach your PRD or relevant documents
5. Let the AI process your request

Method 2: Using the Copy Script

# Make the script executable (first time only)
chmod +x copy-prompt.sh

# List all available prompts
./copy-prompt.sh --list

# Copy a prompt to clipboard
./copy-prompt.sh interactive-prd-creation-prompt.md

Then paste the prompt into your AI assistant and proceed as normal.

Compatibility

These prompts work with any modern LLM, including:

• Claude (Anthropic) - Claude 3.5, Claude 4.x
• GPT-4 / ChatGPT (OpenAI) - GPT-4, GPT-4o, GPT-4.5
• Gemini (Google) - Gemini 1.5, Gemini 2.x
• Local models - Llama, Mistral, etc. (with sufficient context windows)

They also work in AI-powered IDEs like Cursor, Windsurf, Cline, and GitHub Copilot.

Quick Tips

• Provide complete documents when possible
• Answer any clarifying questions the AI asks
• Review and customize AI outputs before implementation
• Use the prompts in sequence for best results
• For complex projects, iterate through prompts as needed

Contributing

See CONTRIBUTING.md for guidelines on submitting new prompts, quality standards, and testing approach.

License

This project is licensed under the MIT License - see the LICENSE file for details.