AI Strategy: How to Choose What AI Product to Implement
创建于 更新于
摘要
The paper introduces the eROI framework to select AI investments by decomposing expected ROI into three independent components: Value if Successful, Likelihood of Success, and Investment Required. This separation avoids the common ROI trap and supports portfolio-level decision making rather than chasing a single top project. It emphasizes enumerating many ideas, rating them across the three pillars with coarse scales (Low–Very High), and constructing a diversified mix of bets. Compass case studies illustrate how projects with high eROI (e.g., Likely-to-Sell recommendations) outperform weaker ideas (e.g., Time-on-Market pricing) and how rigorous consideration of uncertainty guides governance and deployment. The framework also highlights responsible deployment considerations (privacy, fairness, disclosure) as part of the Integration and Productization effort that can influence a project’s eROI. [page::1]
