ai-pm-method-prediction-machines

What problem does it solve?

AI product managers frequently lack a structured, evidence-based economic framework to evaluate AI feature feasibility, commercialization potential, and strategic value, leading to wasted resources on low-impact AI initiatives and missed opportunities for sustainable competitive moats.

Core Features & Use Cases

• P/J/A Task Decomposition Framework: Break down end-to-end workflows into Prediction, Judgment, and Action tasks to identify high-ROI AI automation opportunities and avoid misallocating resources to low-value prediction tasks.
• AI Canvas 7-Grid Template: A pre-PRD evaluation tool that ensures all critical AI feature requirements (prediction targets, judgment trade-offs, feedback loops) are defined before development starts.
• 10 Evidence-Based Decision Rules: Guidelines sourced from the 2022 updated edition of Prediction Machines to guide AI pricing, data moat design, human-AI collaboration, and explicit trade-off choices.
• Use Case: An AI PM evaluating an AI resume polishing tool can use the P/J/A framework to identify that while prediction (resume text generation) is low-cost, the real strategic moat lies in judgment (HR trust in AI-modified resumes), avoiding wasted build effort on a low-value feature.

Quick Start

Use the ai-pm-method-prediction-machines skill to evaluate whether your planned AI feature has a clearly defined prediction target, explicit judgment trade-offs, and a functional feedback loop before committing to development resources.