recommendation-canvas

Recommendation Canvas: AI Product Evaluation and Proposal Canvas

Skill Overview

Recommendation Canvas is a framework that helps product managers systematically assess whether an AI product idea is worth investing in, across dimensions including business outcomes, customer outcomes, problem statements, solution hypotheses, positioning, PESTEL risks, and value justification.

Applicable Scenarios

Leadership has requested an AI feature, and you need to quickly determine whether it is truly worth investing development resources in and provide a go/no-go basis.

You have multiple AI solution options and need to compare them objectively based on business outcomes and risks, rather than simply choosing the newest or most impressive option.

You need to submit a persuasive AI product proposal to executives or decision-makers, explaining its value, assumptions, and risks in order to secure budget and resource support.

Core Functions

Ten-section structured evaluation canvas: Breaks a proposal into ten sections—business outcomes, product outcomes, problem statement, solution hypothesis, positioning statement, assumptions and unknowns, PESTEL risks, value justification, success metrics, and next steps—to produce an executive-ready strategic proposal rather than a feature list.

Hypothesis and validation experiment design: Uses the If/Then format to articulate solution hypotheses and plans lightweight “Tiny Acts of Discovery” and Proof-of-Life validation criteria, treating the solution as a bet to be validated rather than a commitment.

PESTEL risk scanning and SMART metrics: Identifies risks requiring investigation and ongoing monitoring across the six dimensions of political, economic, social, technological, environmental, and legal factors, and defines measurable success metrics using the SMART framework.

Frequently Asked Questions

How is Recommendation Canvas different from a PRD?

It is not a PRD. Recommendation Canvas is a strategic proposal that answers “Why is this worth doing? Which assumptions need to be validated first? How will success be measured?” A PRD, by contrast, is an execution document that describes detailed requirements. The correct sequence is to use the canvas to build the investment case first, then write the PRD after validation is complete.

What should be prepared before using it?

You should first complete some initial discovery: user research, pain-point validation, target audience profiles, the competitive landscape, and business constraints. The canvas synthesizes existing insights into a proposal; it cannot replace research. Without validating the problem first, the output will be based solely on assumptions.

When is it not suitable?

Small feature changes do not require the full canvas, as using it in such cases may lead to over-engineering. It is also unsuitable when no user research has been conducted. In addition, it cannot replace experimentation itself—the canvas tells you what to validate, while validation still requires actual experiments. Completing a full canvas typically takes approximately 30–45 minutes.