Skill: Product management AI

Purpose

Assist with core product management activities including writing product requirements documents (PRDs), analyzing feature requests, synthesizing user research, planning roadmaps, and communicating product decisions to stakeholders and engineering teams.

When to use this skill

• You need to write or update PRDs with clear requirements, success metrics, and technical considerations.
• You're evaluating feature requests and need structured analysis of impact, effort, and priority.
• You need to synthesize user research findings into actionable insights.
• You're planning roadmaps and need to organize, prioritize, and communicate plans.
• You need to communicate product decisions clearly to engineering, design, and business stakeholders.
• You're doing competitive analysis or market research synthesis.
• You need to track and analyze product metrics to inform decisions.

Key capabilities

Unlike point-solution PM tools:

• Integrated with codebase: Can reference actual code, APIs, and technical constraints.
• Context-aware: Understands your specific product, architecture, and technical debt.
• Flexible templates: Adapt documentation to your organization's needs.
• Version controlled: All artifacts live in git alongside code.
• Collaborative: Works within existing dev workflows (PRs, issues, docs).

Inputs

• Product context: Current state, key stakeholders, strategic goals.
• Feature requests: User feedback, business needs, or strategic initiatives.
• Technical constraints: Known limitations, dependencies, or technical debt.
• User research: Interview notes, survey results, analytics data.
• Business goals: Metrics, OKRs, or success criteria to optimize for.

Out of scope

• Making final product decisions (this is the PM's job; the skill assists).
• Managing stakeholder relationships and politics.
• Detailed UI/UX design work (use design tools and collaborate with designers).
• Project management and sprint planning (use project management tools).

Conventions and best practices

PRD structure

A good PRD should include:

1. Problem statement: What user pain point or business need are we addressing?
2. Goals and success metrics: What does success look like quantitatively?
3. User stories and use cases: Who will use this and how?
4. Requirements: Functional and non-functional requirements, prioritized.
5. Technical considerations: Architecture implications, dependencies, constraints.
6. Design and UX notes: Key interaction patterns or design requirements.
7. Risks and mitigations: What could go wrong and how to address it.
8. Launch plan: Rollout strategy, feature flags, monitoring.
9. Open questions: What still needs to be decided or researched.

Feature prioritization

Use structured frameworks to evaluate features:

• RICE: Reach × Impact × Confidence / Effort
• ICE: Impact × Confidence × Ease
• Value vs. Effort: 2×2 matrix plotting value against implementation cost
• Kano Model: Categorize features into basic, performance, and delighters

User research synthesis

When synthesizing research:

1. Identify patterns: What themes emerge across participants?
2. Quote verbatim: Include actual user quotes to illustrate points.
3. Quantify when possible: "7 out of 10 participants said..."
4. Segment findings: Different user types may have different needs.
5. Connect to metrics: How do qualitative findings explain quantitative data?

Roadmap planning

Effective roadmaps should:

• Theme-based: Group work into strategic themes, not just feature lists.
• Time-horizoned: Now / Next / Later or Quarterly structure.
• Outcome-focused: Emphasize goals and outcomes, not just outputs.
• Flexible: Leave room for learning and adjustment.
• Communicated clearly: Different views for different audiences.

Required behavior

1. Understand context deeply: Review existing docs, code, and prior discussions before proposing changes.
2. Ask clarifying questions: Don't assume; clarify ambiguous requirements or goals.
3. Be specific and actionable: Avoid vague language; provide concrete, testable requirements.
4. Consider tradeoffs: Explicitly discuss pros/cons of different approaches.
5. Connect to strategy: Tie features and decisions back to higher-level goals.
6. Involve stakeholders: Identify who needs to review or approve.
7. Think through edge cases: Don't just focus on happy paths.
8. Make it measurable: Propose concrete metrics to track success.

Required artifacts

Depending on the task, generate:

• PRD document: Comprehensive product requirements in markdown format.
• Feature analysis: Structured evaluation of a feature request.
• Research synthesis: Summary of user research findings with insights.
• Roadmap document: Organized view of planned work with themes and timelines.
• Decision document: Record of key product decisions and rationale.
• Competitive analysis: Comparison of competitor features and approaches.
• Metric definitions: Clear definitions of success metrics and how to measure them.

Implementation checklist

Writing a PRD

• Understand the problem space and strategic context
• Review related code, APIs, and technical constraints
• Interview key stakeholders (engineering, design, business)
• Research user needs and competitive landscape
• Draft problem statement and goals
• Define user stories and use cases
• Specify functional and non-functional requirements
• Document technical considerations and dependencies
• Define success metrics and measurement approach
• Identify risks and mitigation strategies
• Plan rollout and launch approach
• Review with stakeholders and iterate

Analyzing a feature request

• Clarify the user problem or business need
• Identify target users and use cases
• Estimate impact (users affected, business value)
• Assess implementation effort and complexity
• Identify dependencies and risks
• Check alignment with product strategy
• Compare against alternatives
• Calculate prioritization score (RICE, ICE, etc.)
• Make recommendation with clear reasoning

Synthesizing user research

• Review all research materials (transcripts, notes, data)
• Identify key themes and patterns
• Extract representative quotes
• Segment findings by user type if relevant
• Connect qualitative findings to quantitative data
• Formulate insights and implications
• Generate actionable recommendations
• Prioritize recommendations by impact

Planning a roadmap

• Review strategic goals and OKRs
• Collect input from stakeholders
• Assess current state and technical debt
• Group potential work into strategic themes
• Prioritize themes and initiatives
• Estimate sizing and dependencies
• Organize into time horizons (Now/Next/Later)
• Define success criteria for each initiative
• Create views for different audiences
• Review and socialize with stakeholders

Example workflows

Example 1: Writing a PRD for a new feature

# PRD: Advanced Search Functionality

## Problem Statement
Users frequently report difficulty finding specific items in our catalog when they have multiple criteria (price range, location, category, features). Our current search only supports simple text queries, leading to:
- High bounce rates on search results pages (65% bounce rate vs 32% site average)
- Increased support tickets asking for search help (150/month)
- Lost conversion opportunities (estimated $500K annual revenue impact)

## Goals and Success Metrics
**Primary Goal**: Enable users to find relevant items quickly using multiple filters.

**Success Metrics**:
- Reduce search result page bounce rate from 65% t