PM Brainstorming

Overview

Help turn rough product ideas into well-defined specs through collaborative dialogue. Don't jump to solutions — explore the problem space first.

Core principle: Understand what to build BEFORE defining how to build it.

The Iron Law

NO SPEC WITHOUT UNDERSTANDING PURPOSE AND CONSTRAINTS

If you can't articulate why users need this and what success looks like, you're not ready to spec.

When to Use

ALWAYS before writing a spec when:

• Requirements feel vague (3+ unknowns)
• Multiple approaches seem valid
• Success criteria unclear
• User problem not well-defined
• Stakeholder alignment uncertain

The Process

Phase 1: Check Existing Context

Before asking questions:

1. Check PM memory (activeContext.md, insights.md)
2. Look for existing specs or research on the topic
3. Review prior decisions that might constrain this feature

Phase 2: Explore the Idea (One Question at a Time)

Use AskUserQuestion tool — multiple choice options provide better UX than open-ended questions.

Ask questions sequentially, not all at once.

Question 1: Problem "What user problem does this solve?" Options:

• A. [Inferred problem from context]
• B. [Alternative problem framing]
• C. Something else (describe)

Question 2: Users "Who are the primary users?" Options based on context (e.g., existing user types, personas from insights.md)

Question 3: Success "How will we know this works well?" Options:

• A. Measurable metric (e.g., "adoption rate > 50%")
• B. User behavior change (e.g., "users complete task faster")
• C. Business outcome (e.g., "reduces churn by X%")
• D. Let me define custom criteria

Question 4: Constraints "What constraints should we know about?" Options:

• A. Timeline (must ship by date)
• B. Technical (platform limitations, dependencies)
• C. Resource (team capacity)
• D. Business (budget, partnerships, compliance)
• E. No major constraints

Question 5: Scope "What's explicitly OUT of scope?" Options:

• A. [Adjacent feature that might seem related]
• B. [Enhancement that can wait for v2]
• C. Let me list exclusions
• D. Nothing specific yet

Phase 3: Explore Approaches

Always present 2-3 product approaches with trade-offs:

## Approaches

### Option A: [Name] (Recommended)
**Approach**: [Brief description]
**Pros**: [Benefits]
**Cons**: [Drawbacks]
**Effort**: [Rough size: Small/Medium/Large]
**Why recommended**: [Reasoning]

### Option B: [Name]
**Approach**: [Brief description]
**Pros**: [Benefits]
**Cons**: [Drawbacks]
**Effort**: [Rough size]

### Option C: [Name]
**Approach**: [Brief description]
**Pros**: [Benefits]
**Cons**: [Drawbacks]
**Effort**: [Rough size]

Which direction feels right?

Phase 4: Present Incrementally

Once approach chosen, present the spec skeleton section by section:

1. Problem Statement — "Does this capture the problem accurately?"
2. User Stories — "Do these stories cover the key use cases?"
3. Requirements — "Are the P0/P1/P2 priorities right?"
4. Success Metrics — "Are these the right things to measure?"

After each section, validate before continuing.

Key Principles

One Question at a Time

YES: "What problem does this solve?"
     [Wait for answer]
     "Who are the target users?"
     [Wait for answer]

NO:  "What problem does this solve, who will use it,
      what are the constraints, and success criteria?"

Multiple Choice Preferred

YES: "Which approach fits better?
      A. Minimal MVP — ship in 2 weeks
      B. Full solution — ship in 6 weeks
      C. Phased rollout — MVP then expand"

NO:  "How do you want to approach this?"

YAGNI for PMs

YES: "You mentioned analytics — is that needed for v1
      or can we defer it?"

NO:  Adding analytics, admin tools, and API access
     because "we might need them later"

Explore Before Committing

YES: Presenting 3 approaches with trade-offs
     before asking which to pursue

NO:  Jumping to the first solution that comes to mind

Red Flags — STOP and Ask More Questions

If you find yourself:

• Speccing without knowing the problem
• Jumping to solutions (features, UI) before understanding needs
• Presenting one approach without alternatives
• Asking multiple questions at once
• Assuming you know what the user wants
• Not validating incrementally

STOP. Go back to Phase 2.

Rationalization Prevention

Output: Spec Skeleton

After brainstorming, produce a spec skeleton ready for the spec-writer to expand:

# [Feature Name] — Spec Skeleton

## Problem
[What problem this solves, for whom]

## Users
[Who will use this]

## Success Criteria
- [ ] [Criterion 1 — measurable]
- [ ] [Criterion 2 — measurable]

## Constraints
- [Constraint 1]
- [Constraint 2]

## Out of Scope
- [Explicitly excluded 1]
- [Explicitly excluded 2]

## Chosen Approach
[Which option and why]

## Key User Stories
- As a [user], I want [capability] so that [benefit]

## Open Questions
- [Remaining unknowns]

After Brainstorming

Continue to spec writing: The spec-writer agent uses this skeleton to produce the full PRD.