| name | start-2-1 |
| description | Module 2.1: Write a PRD - Partner with AI to write better PRDs using templates, context, and multi-perspective review. Use when the student types /start-2-1.
|
| disable-model-invocation | true |
Setup
Read .cursor/SCRIPT_INSTRUCTIONS.md and follow it for everything below.
ACTION: Silently stage bundled lesson assets without overwriting student work.
cp -rn .cursor/skills/start-2-1/assets/* . 2>/dev/null || true
Module 2.1: Write a PRD - Teaching Script
Introduction & Overview
- Welcome to Module 2.1: Write a PRD!
- This module teaches you how to partner with AI to write better Product Requirements Documents faster
- The key insight: AI isn't here to ghostwrite your PRD - it's here to help you think more clearly and move faster
- You'll drive every decision, I'll help you explore options and refine your thinking
- STOP: Ready to dive in?
- USER: Confirms ready
Show Complete Workflow
- Let me show you the complete workflow we'll follow today
- ACTION: Display Mermaid diagram showing the PRD writing workflow from start to finish
graph TD
A[Phase 1: Choose Foundation] --> B[Phase 2: Build Context]
B --> C[Phase 3: Sharpen Thinking]
C --> D[Phase 4: Generate Options]
D --> E[Phase 5: Multi-Perspective Review]
E --> F[Phase 6: Address Feedback]
F --> G[✅ Completed PRD]
style A fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style B fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style C fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style D fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style E fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style F fill:#4A90E2,stroke:#2E5C8A,stroke-width:2px,color:#fff
style G fill:#50C878,stroke:#2D7A4A,stroke-width:3px,color:#fff
- Here's what each phase does:
- Phase 1: Choose Foundation (select your PRD template)
- Phase 2: Build Context (company background + user research)
- Phase 3: Sharpen Thinking (Socratic questioning to refine your idea)
- Phase 4: Generate Options (create 3 different PRD versions)
- Phase 5: Multi-Perspective Review (get feedback from 3 reviewer personas)
- Phase 6: Address Feedback (make strategic decisions and implement changes)
- Result: Completed PRD ready to share with stakeholders
- This gives you a complete map of where we're going
- STOP: Any questions before we begin?
- USER: Confirms
File Orientation
- Perfect! Now let me walk you through the files we'll be using today
- ACTION: List files in lesson-modules/2.1-write-prd/ folder structure
- Here's what each file does:
- prd-templates/: Two professional templates you can choose from (Carl's detailed version, Lenny's 7-question version)
- taskflow-company-context.md: Everything about the company (product, customers, strategy)
- user-research/pain-points.md: Real user research insights to ground your PRD
- socratic-questioning.md: Framework for sharpening your feature thinking
- reviewers/: Three different personas who'll review your PRD (engineer, executive, user researcher)
- STOP: Take a minute to look at the folder structure - do you see all these files?
- USER: Confirms
Phase 1: Choose Your Foundation
- Great! Let's start Phase 1: Choose Your Foundation
- You need a PRD template to work from
- I have two professional templates you can use, or you can bring your own
- ACTION: Read both prd-templates/Carls-PRD-Template.md and prd-templates/Lennys-PRD-Template.md
- Here are your options:
- Carl's PRD Template: Comprehensive structure with Problem Alignment, Solution Alignment, and Development Planning - great for complex features needing stakeholder buy-in
- Lenny's PRD Template: Minimal 7 questions (What, Why, Who, Success, etc.) - perfect for moving fast on smaller features
- Your Own Template: For real PM work, you'd use whatever template your company prefers - these are just examples
- STOP: Which would you like to use for today's exercise? Say: "Carl's template", "Lenny's template", or "I'll use my own"
- USER: Chooses template
Introduce the Feature
- Excellent choice!
- For today's practice, we'll build a PRD for this feature: An AI voice chat interface for managing your to-do list
- Think of it like talking to an assistant who helps you organize tasks
- Now let's move to Phase 2: Build Context
- Great PRDs are grounded in company strategy and user research
- Before we write anything, I need context on TaskFlow and what users are asking for
- STOP: Now say: "Review taskflow-company-context.md and pain-points.md before we begin"
- USER: Review taskflow-company-context.md and pain-points.md before we begin
Phase 2: Build Context
- ACTION: Read taskflow-company-context.md and user-research/pain-points.md
- Perfect! I've read both files and now have full context on:
- TaskFlow's product (project management for remote teams, 10K users, $2.5M ARR)
- Your target customers (SMBs with 5-20 person teams)
- Key pain points from user research (complexity overwhelm, need for quick task capture)
- This context will make the PRD much stronger
- STOP: Make sense so far?
- USER: Confirms
Phase 3: Sharpen Your Thinking
- Now for Phase 3: Sharpen Your Thinking
- Before we write anything, let's refine the feature idea through strategic questions
- This is where AI partnership really shines - I can help you think more rigorously
- I'm going to use the Socratic questioning framework from socratic-questioning.md
- I'll ask you 3 strategic questions about your AI voice chat feature
- For this demo, I'll give you multiple-choice options to keep things moving
- In real PM work, the AI could ask more open-ended questions and you'd type out your thinking
- But multiple choice makes this teaching module faster and easier
- STOP: Now say: "Use the Socratic questioning framework to sharpen my thinking on this feature"
- USER: Use the Socratic questioning framework to sharpen my thinking on this feature
Socratic Question 1
- ACTION: Read socratic-questioning.md
- Great! I've loaded the framework - let's dive into the questions
- Question 1: What's the core job-to-be-done that voice chat solves for TaskFlow users?
- STOP: What's your thinking?
- A) Hands-free task capture while multitasking (meeting, driving, cooking)
- B) Faster than typing for quick tasks
- C) More natural/conversational interface feels less intimidating
- D) Accessibility for users who struggle with typing
- USER: Chooses A, B, C, or D
Socratic Question 2
- Great insight! That's a solid foundation for the feature
- Question 2: How does this fit TaskFlow's strategic focus on remote teams?
- STOP: Your thoughts?
- A) Remote workers are often in meetings where typing is disruptive but voice notes work
- B) Distributed teams need async communication - voice memos are async
- C) Remote workers multitask more (kids, errands) - need hands-free tools
- D) Doesn't really fit - this is just a cool feature
- USER: Chooses A, B, C, or D
Socratic Question 3
- Exactly! You're seeing the strategic connection to TaskFlow's positioning
- Question 3: What's the main risk or concern with adding voice chat to a productivity tool?
- STOP: What worries you?
- A) Privacy/security - users won't want voice data stored
- B) Accuracy - voice recognition errors could frustrate users
- C) Context mismatch - many users work in open offices or quiet spaces where they can't talk
- D) Complexity - adds cognitive load to learn a new interaction mode
- USER: Chooses A, B, C, or D
Transition to Phase 4
- That's an important consideration to address in the PRD!
- Excellent thinking! Your feature idea is getting sharper
- I can see this taking shape: Voice chat for quick task capture, especially valuable for remote workers who need hands-free interaction
- Now we're ready for Phase 4: Generate Options
- Here's a pro tip: AI is non-deterministic - the same inputs can produce different outputs
- Different language, different emphasis, different structure - even with identical starting conditions
- Let's generate 3 draft PRDs using identical inputs (your template + context + research + answers)
- This gives you options to compare - pick whichever framing resonates most, or cherry-pick the best parts from each
- STOP: Now say: "Generate 3 draft PRDs"
- USER: Generate 3 draft PRDs
Phase 4: Generate Three PRD Versions
- ACTION: Create ai-chat-todo-prd-v1.md
- Use chosen template structure
- Incorporate company context, user research, and Socratic question answers
- ACTION: Create ai-chat-todo-prd-v2.md
- Same inputs as v1 (template + context + research + answers)
- AI will naturally produce variations in language and emphasis
- ACTION: Create ai-chat-todo-prd-v3.md
- Same inputs as v1 and v2
- Another variation to compare
- Done! I've created 3 draft PRDs using identical inputs
- Each one might emphasize different aspects, use different language, or structure arguments differently
- That's the nature of AI - non-deterministic outputs from the same inputs
- STOP: Review all 3 versions - which one resonates most? Say "v1", "v2", "v3", or "I want to combine elements from multiple versions"
- USER: Chooses version or combines
Phase 5: Multi-Perspective Review
- Good eye! You picked the version that works best
- Now let's get this PRD stress-tested from multiple perspectives
- Now for Phase 5: Multi-Perspective Review
- Before anyone sees your PRD, let's get feedback from three different perspectives
- I have three reviewer personas in the reviewers/ folder: engineer (technical feasibility), executive (business value), user researcher (UX/needs)
- In real PM work, you could create your own reviewer personas - maybe a designer, data scientist, support lead - whatever perspectives matter for your product
- These are reusable files you can refine over time, just like we talked about with templates and workflows earlier
- I'll review your PRD from all three perspectives and consolidate the feedback into one document
- STOP: Now ask me for a review from all the reviwers, say: "Review this PRD from all three perspectives in @reviewers/ and consolidate the feedback into a new doc"
- USER: Review this PRD from all three perspectives in @reviewers/ and consolidate the feedback into a new doc
Execute Multi-Perspective Review
- ACTION: Read all files in reviewers/ folder (engineer.md, executive.md, user-researcher.md)
- ACTION: Read the chosen PRD version
- ACTION: Create ai-chat-todo-prd-review.md consolidating feedback from all three perspectives
- Engineering Perspective: Technical feasibility, implementation complexity, infrastructure needs, potential issues, recommendations
- Executive Perspective: Business value, strategic fit, market differentiation, resource commitment, ROI concerns, recommendations
- User Research Perspective: User needs validation, usability concerns, accessibility, adoption risks, recommendations
- Common Themes: Identify patterns across all three perspectives
- Priority Feedback Items: Rank by importance
- Done! I've reviewed your PRD from all three perspectives and created a comprehensive consolidated review in ai-chat-todo-prd-review.md
- STOP: Take a look at the review document, then say: "Help me address the feedback here"
- USER: Views document, then says: Help me address the feedback here
Phase 6: Address Feedback - Item 1
- Perfect! Phase 6: Address the Feedback
- Let's work through the major feedback items and improve your PRD
- I'll present each key piece of feedback as a multiple-choice question - you choose the approach, I implement it
- Looking at the engineering feedback: "Start with text-based version first, add voice in v2 - reduces risk"
- STOP: How should we address this?
- A) Agree - phase the rollout with text first
- B) Disagree - ship voice from day 1 since it's the core value prop
- C) Hybrid - build both but make voice optional beta feature
- USER: Chooses A, B, or C
Address Feedback - Item 2
- Smart choice! That balances risk with impact
- Next feedback from executive: "Add monetization angle - could this be a premium feature?"
- STOP: How should we address this?
- A) Yes - make it Premium/Max tier only
- B) No - free feature to drive engagement
- C) Freemium - basic voice free, advanced features (transcripts, voice commands) premium
- USER: Chooses A, B, or C
Address Feedback - Item 3
- Good thinking on the monetization strategy!
- Final feedback from UX research: "Strong onboarding flow needed - users might not understand how to use voice"
- STOP: How should we address this?
- A) Guided tutorial on first use with voice prompts
- B) Subtle hints/tooltips that appear contextually
- C) Video demo in product tour
- USER: Chooses A, B, or C
Implement Feedback Changes
- Excellent! That onboarding approach should help with adoption
- Perfect! You've made strategic decisions on all the major feedback
- Now let me implement these changes directly in your PRD
- ACTION: Edit the chosen PRD version (v1, v2, or v3) to incorporate feedback decisions
- Update phasing approach based on engineering feedback choice
- Add monetization section based on executive feedback choice
- Add onboarding/UX section based on UX feedback choice
- Refine success metrics to reflect new approach
- Update timeline if needed
- Done! I've updated your PRD with all the feedback responses
- You should see the changes highlighted in the file - this is how you'd actually iterate on a PRD in real work
- It now includes the phasing strategy, monetization approach, and UX onboarding plan you decided on
- STOP: Review the updated PRD file - do you see all the changes incorporated?
- USER: Confirms
Recap Accomplishments
- Excellent work! Let's recap what you just accomplished:
- ✅ Chose a professional PRD template (or brought your own)
- ✅ Grounded your thinking in company context and user research
- ✅ Sharpened your feature idea through Socratic questioning
- ✅ Generated 3 draft PRDs to compare
- ✅ Got multi-perspective feedback (engineering, executive, UX)
- ✅ Made strategic decisions on how to address feedback
- ✅ Created a production-ready PRD incorporating all insights
- Here's the key insight: YOU drove this entire process
- I didn't ghostwrite a PRD for you
- STOP: Do you see how you made all the strategic decisions while I helped you think through options?
- USER: Confirms understanding
Key Insights & Customization
- This is what AI partnership looks like for PMs:
- AI helps you think more clearly (Socratic questions)
- AI helps you explore options (multiple drafts)
- AI helps you stress-test ideas (multi-perspective review)
- But YOU make all the decisions
- Now here's the best part: This workflow is YOURS to customize
- You can modify any step, add new steps, or remove steps that don't work for you
- Want to add a fourth reviewer perspective? Create a designer.md file in the reviewers folder
- Don't like Socratic questions? Skip that phase or replace it with a different framework
- Want to use examples? Add a prd-examples/ folder with great PRDs from PMs at your company and @ mention them
- Prefer different templates? Swap them out
- This is just ONE way to approach PRD writing - you'll build YOUR workflow over time
- The templates, context files, reviewer personas, and frameworks are all living documents you refine as you learn what works
- Congratulations - you now know how to partner with AI to write better PRDs faster!
- Next up: Module 2.2 is all about data analysis
- You'll learn how to analyze funnels, build ROI models, and run experiment analysis - the complete data-driven PM workflow
- It's incredibly practical and you'll create real analysis documents you could present to leadership
- STOP: When you're ready, type /start-2-2 to begin Module 2.2: Analyze Data
- USER: Acknowledges
END OF MODULE 2.1
ACTION: Before wrapping up, record this lesson as complete by running this WITHOUT NARRATING the raw output:
fspm progress complete cursor-pms-2-1
If it fails because the fspm CLI isn't installed, follow the teaching rules' missing-CLI guidance: tell the learner progress tracking needs the FSPM CLI, offer to install it, and continue the wrap-up either way.
Then close out with the student's options, in natural language (never as a command list):
- Mention they can leave feedback on this lesson anytime — if they have some, collect it conversationally and submit it with the CLI (see teaching rules).
- Remind them, briefly and only if it fits the moment, that they can always ask for a recap, a quiz on what they just did, a saved note, or where they stand in the course.
- End with the next step: when you're ready for the next lesson, use
/clear first, then:
/start-2-2
