recommendation-canvas

This skill provides a structured canvas for evaluating and proposing AI product solutions, helping product managers build a defensible recommendation for stakeholders. It solves the problem of pitching AI features that carry higher uncertainty by forcing explicit articulation of why the solution is worth building, what assumptions must be validated, and how success will be measured, rather than jumping straight to a feature spec.

The canvas synthesizes several PM frameworks into one strategic view with components including business outcome, product/customer outcome, persona-centric problem statement, if/then solution hypothesis with experiments, positioning statement, and assumptions and risks with value justification. It composes with sibling skills, drawing the problem-framing narrative from a problem-statement skill and the epic hypothesis format from an epic-hypothesis skill, and ships with a template and a worked sample. Quality checks throughout emphasize customer-centric, outcome-oriented, and research-validated framing over feature-first language.

It is aimed at product managers preparing go/no-go recommendations for decision-makers, particularly for AI-powered features where risk and uncertainty are elevated. It was created for Dean Peters' Productside 'AI Innovation for Product Managers' class. As a component-type skill it produces a written strategic proposal and involves no code execution.