Synthesize

---
name: wicked-garden-product-synthesize
description: |
  Generate actionable recommendations from customer voice insights. This skill
  should be used when the user needs to translate analyzed feedback into product
  priorities, feature recommendations, or strategic guidance.

  Use when: "recommendations from feedback", "translate feedback to priorities",
  "feature priorities from customer data", "synthesize feedback into action"
metadata:
  role: module
  phases: "clarify,design,review"
  archetypes: "*"
---

# Synthesize Skill

Transform customer voice analysis into actionable product recommendations.

> Tier-3 knowledge module behind the **`synthesize`** action of the
> `wicked-garden-product` domain skill (`skills/product/SKILL.md`). Canonical
> invocation: `/wicked-garden-product synthesize [args]` or plain natural
> language — the examples below use that form.

## When to Use

- After analyzing customer feedback
- Product planning or roadmap prioritization
- Feature decision support
- User asks "what should we build based on customer feedback"

## Usage

```bash
# Synthesize all recent insights
/wicked-garden-product synthesize

# Prioritize by impact level
/wicked-garden-product synthesize --priority high

# Feature-specific synthesis
/wicked-garden-product synthesize --feature "mobile-app"

# Brief format (for quick decisions)
/wicked-garden-product synthesize --format brief
```

## Synthesis Framework

### 1. Insight Extraction
Identify: Pain Points, Opportunities, Risks, Strengths

### 2. Recommendation Generation
Score = (Customer_Value × Urgency) / Effort
See [refs/prioritization.md](refs/prioritization.md) for detailed scoring.

### 3. Customer Journey Mapping
Map feedback across: Awareness → Evaluation → Onboarding → Adoption → Retention → Advocacy
See [refs/journey-mapping.md](refs/journey-mapping.md) for stage analysis.

### 4. Priority Levels
- **P0**: Churn risk, blocker, security
- **P1**: Major friction, competitive gap
- **P2**: Enhancement, efficiency
- **P3**: Nice-to-have, edge case

## Output Format

```markdown
## Customer Voice Synthesis: {Context}

### Executive Summary
{2-3 sentences on what customers need most}

### Top Recommendations
1. **{Recommendation}** - {Priority}
   - Impact: {HIGH/MEDIUM/LOW}
   - Evidence: {N} mentions, {sentiment}, {trend}
   - Action: {specific what to do}
   - Effort: {LOW/MEDIUM/HIGH}

{Top 3-5 recommendations}

### Customer Journey Insights
- **Onboarding**: {key pain point or win}
- **Adoption**: {key pain point or win}
- **Retention**: {key pain point or win}

### Strategic Considerations
- **Strengths**: {what to maintain}
- **Risks**: {what threatens retention}
- **Opportunities**: {what could differentiate}

### Quick Wins
{1-3 low-effort, high-impact improvements}

### Long-term Bets
{1-2 strategic investments based on customer needs}
```

## Synthesis Process

1. Load theme and sentiment analysis from the product voice analysis directory (resolved via `resolve_path.py wicked-garden:product`)
2. Sort themes by priority score, filter by threshold
3. Generate recommendations: theme → action, score by impact/effort
4. Map to customer journey stages, identify friction points
5. Format output: executive summary first, top 3-5 recommendations

## Integration

**wicked-crew**: Auto-inject synthesis into `product:requirements:started` events
**wicked-garden-mem (store action)**: Store high-priority insights as decision memories
**Native tasks**: Link recommendations to roadmap tasks via TaskCreate with `metadata.event_type="task"`

## Rules

- Maximum 5 recommendations (context efficiency)
- Always include effort estimates
- Provide confidence levels (HIGH/MEDIUM/LOW)
- Cite supporting evidence (N mentions, sentiment)
- Distinguish quick wins from strategic bets
- Keep output under 1000 words
- Be honest about unknowns and limitations