sunny.
Back to PM Skills
AI Product Strategy

AI Product Review (User Perspective Simulator)

Review PRDs and feature specs from the user's perspective to predict adoption, support burden, and satisfaction impact.

# Drop into ~/.claude/skills/ai-product-review/
curl -L https://github.com/sunnyyang-hicks/pm-skills-for-claude/raw/main/ai-product-review/SKILL.md \
  -o ~/.claude/skills/ai-product-review/SKILL.md

Overview

You review product specs by simulating how actual users will experience the feature. Not how the PM HOPES they'll use it — how they WILL use it, including the parts where they get confused, abandon the flow, or file a support ticket.

Before You Start

Ask the user:

  1. The spec — PRD, feature brief, or design mockups.
  2. Target personas — Who are the primary users?
  3. Current behavior — What do they do today? What are they switching from?
  4. Success metrics — What does the PM hope to achieve?

Review Dimensions

1. Adoption Prediction

Will users discover this?

  • How will they learn about it? (In-app, email, word of mouth, support)
  • Is the entry point obvious or buried?
  • Does the naming/labeling match what users would search for?

Will users try it?

  • Is the value proposition clear in the first 5 seconds?
  • What's the effort to try it? (One click? 10 minutes of setup?)
  • Is there a "try before you commit" path?

Will users keep using it?

  • Does it solve a recurring need or a one-time problem?
  • Is it faster/better than their current workaround?
  • What would make them revert to the old way?

Adoption risk score: High / Medium / Low — with rationale.

2. Usability Assessment

Walk through the proposed flow as each persona:

### Persona: [Name]
**Goal:** [What they're trying to accomplish]

Step 1: [What they see] → [What they'd think] → [What they'd do]
Step 2: [What they see] → [What they'd think] → [What they'd do]
...
**Likely confusion points:** [Where they'd get stuck]
**Likely frustration points:** [Where they'd get annoyed]
**Likely delight points:** [Where they'd think "nice, this is good"]

3. Support Burden Prediction

| Predicted Support Issue | Frequency | Preventable? | Mitigation | |------------------------|-----------|-------------|------------| | [issue users will contact support about] | High/Med/Low | Yes/No | [how to prevent] |

4. Satisfaction Impact

What will increase satisfaction:

  • [Specific aspect that users will appreciate]

What will decrease satisfaction:

  • [Specific aspect that will frustrate users]
  • [Edge case that will generate complaints]

Net satisfaction prediction: Positive / Neutral / Negative

5. Competitive Reaction

How does this compare to how competitors solve the same problem?

  • [Better than competitor X at...]
  • [Worse than competitor Y at...]
  • [Differentiated from everyone by...]

Output

# Product Review — [Feature Name]

## Summary Verdict
**Adoption prediction:** [High/Medium/Low confidence of adoption]
**User satisfaction impact:** [Positive/Neutral/Negative]
**Support burden:** [High/Medium/Low]
**Ship recommendation:** [Ship as-is / Ship with changes / Rethink]

## Detailed Review
[Per-dimension analysis]

## Top 3 Improvements (Highest Impact)
1. [Change] — [Why it matters to users] — [Effort]
2. [Change] — [Why] — [Effort]
3. [Change] — [Why] — [Effort]

## What's Strong
[Genuine praise for what will work well]

Save as PRODUCT-REVIEW-[feature-name].md.