Recommendation Canvas (AI Product)
10-component canvas for evaluating AI product proposals — business + customer outcomes, hypothesis with tiny experiments, PESTEL risks, value justification, SMART metrics, and next steps.
Your AI feature pitch is "we'll add an AI assistant." Your CFO is asking what it's worth and what could go wrong. The Recommendation Canvas gives both: business outcome ("reduce churn 25%"), customer outcome ("save 5 hours/month"), validated hypothesis, PESTEL risks (yes, including GDPR), SMART metrics, and what to test before committing engineering. Decision-grade, not slide-deck-grade.
Who it's for: PMs proposing AI features, founders pitching AI products to investors, product leaders evaluating AI investments, exec teams making go/no-go calls on high-uncertainty bets
Example
"Build a Recommendation Canvas for AI invoice reminders" → Business Outcome (20% MRR from freelancers) + Hypothesis ("If we provide AI reminders for freelance designers, then we reduce follow-up time 70%") + PESTEL (GDPR for email patterns) + 3 experiments + SMART metrics + 5 next steps
CLAUDE.md Template
New here? 3-minute setup guide → | Already set up? Copy the template below.
npx ccplaybooks add recommendation-canvas# Recommendation Canvas (AI Product Decisions)
Evaluate and propose AI product solutions using a structured canvas: business outcomes, customer outcomes, problem framing, solution hypotheses, positioning, risks, value justification. Built for high-uncertainty AI features that need defensible recommendations.
Not a feature spec. A strategic proposal articulating *why* this is worth building, *what* assumptions need validating, and *how* you'll measure success.
## The 10 Canvas Components
1. **Business Outcome** — what's in it for the business?
2. **Product Outcome** — what's in it for the customer?
3. **Problem Statement** — persona-centric problem framing
4. **Solution Hypothesis** — if/then with experimentsGet new playbooks like this one
One email a week with new Claude Code workflows. Free, like everything here.
No spam. Unsubscribe anytime.
What This Does
Walks through 10 sections (business outcome → product outcome → problem → hypothesis → positioning → assumptions → PESTEL → value justification → SMART metrics → next steps). Built specifically for AI features where uncertainty is high and execs want explicit risk framing.
Pairs with problem-statement, epic-hypothesis, positioning-statement, proto-persona, and pestel-analysis.
Quick Start
mkdir -p ~/Documents/RecommendationCanvas
mv ~/Downloads/CLAUDE.md ~/Documents/RecommendationCanvas/
cd ~/Documents/RecommendationCanvas
claude
Provide problem context, target persona, market intel, and business constraints. Claude drafts the full 10-section canvas with PESTEL risks and SMART metrics.
The 10 Sections
| # | Section | Output |
|---|---|---|
| 1 | Business Outcome | [Direction] [Metric] [Outcome] [Context] |
| 2 | Product Outcome | Customer POV outcome |
| 3 | Problem Statement | Persona narrative |
| 4 | Solution Hypothesis | If/Then + experiments + Proof-of-Life |
| 5 | Positioning | Value prop + differentiation |
| 6 | Assumptions & Unknowns | Explicit list to validate |
| 7 | PESTEL Risks | Investigate + Monitor |
| 8 | Value Justification | C-level convincing |
| 9 | Success Metrics | SMART |
| 10 | What's Next | 5 strategic steps |
Tips & Best Practices
- Business + customer outcomes both required. One without the other is incomplete.
- Hypothesis = bet, not commitment. Frame as "if we... then we will..." with experiments to validate.
- PESTEL specificity matters. "Regulations might change" is theater. "GDPR for storing email timing patterns requires legal review" is actionable.
- Tiny Acts of Discovery before engineering. 2-week prototype + 10 users beats 3-month build for a hypothesis.
- Proof-of-Life criteria upfront. Define quant + qual signals before testing, not after.
Common Pitfalls
- Vague outcomes ("improve UX," "increase revenue")
- Solution-first problem framing ("we need AI X")
- Skipping Tiny Acts of Discovery (hypothesis straight to roadmap)
- Generic PESTEL risks ("regulations might change")
- Value justification without data or dollar amounts