Prompt Engineering for Product Managers
Learn the prompting techniques that produce genuinely useful AI output for PM tasks — and why vague prompts produce vague results.
In practice: Discovery synthesis: 2 weeks → hours
Step 1 of 4
Read
Why Prompting Matters for PMs
Product managers are professional communicators. The skills that make you good at writing a brief — specificity, context, clear desired outcome — are exactly the skills that make you good at prompting AI. The difference is that with AI, you get the output immediately, so you can iterate in real time.
The PM Prompt Framework
Every good PM AI prompt includes:
-
Role context. What kind of PM are you? What type of product? B2B or B2C? What's the stage? "You are helping a B2B SaaS product manager at a 200-person company..."
-
Task and desired output. What do you want? In what format? "...write a one-page PRD for a new notifications feature. Format: background, problem statement, goals, user stories (3–5), non-goals, success metrics."
-
Relevant context. What does the AI need to know to do the task well? "...The target user is an operations manager who receives email digests today and wants real-time alerts for critical events. The feature should not require app download."
-
Constraints. What limitations apply? "...Maximum 500 words. Avoid implementation details. Frame success metrics as user behaviours, not technical metrics."
Prompt Patterns for Common PM Tasks
User story prompt: "Write 5 user stories for [feature] following the format: As a [persona], I want to [action] so that [benefit]. The persona is [describe]. The product is [describe]. Acceptance criteria for each story."
Competitor analysis prompt: "Analyse [competitor product] from the perspective of a PM at a competing company. Cover: target user, core value proposition, key differentiators, apparent weaknesses, and strategic positioning. Base your analysis on publicly available information."
Problem statement prompt: "I'm going to describe a user problem. Help me reframe it as a clear, testable problem statement that separates the symptom from the root cause. User description: [paste interview excerpt or your summary]."
The Iteration Principle
A prompt is rarely perfect on first try. Treat prompting like product iteration: ship something, see what's wrong, refine. "Make it more specific to B2B users" or "Reframe the problem statement to focus on the outcome, not the feature" are valid iterations.
The prompt
Claude's response
Quick check
You ask AI to write a PRD for a new feature. What should you provide for the most useful output?
Key Takeaways
- 1PM communication skills translate directly into prompting skills — specificity and context determine output quality
- 2The four elements of every good PM prompt: role context, task and format, relevant context, and constraints
- 3Build a personal prompt library for your most common tasks (PRDs, user stories, research synthesis)
- 4Treat prompting iteratively — refine based on output quality, just like a product
- 5Vague prompts produce generic output; specific prompts produce immediately useful output
Your challenge this week
Apply what you learned in a real task
Take one feature you're currently scoping or have recently shipped. Use AI to write the user stories or rewrite existing ones with specific acceptance criteria. Then review them with an engineer and note where the AI output needed the most correction.
Before you practise
Think about the last feature that shipped differently than the PRD described. At which point did the misalignment happen — writing the requirements, sharing them with the team, or during implementation? Where could clearer AI-assisted documentation have helped?
Next step
Put it into practice
You've read the lesson — now apply it in a guided hands-on exercise. It takes about 5 minutes.