Information systems strategy in the age of machine learning

Organizations implementing machine learning (ML) often fall into a strategy trap: prioritizing compliance and competitive imitation over differentiation. While viability requires ML capabilities above a certain threshold, the resulting table stakes applications – such as chatbots for ordering and service provision – may meet expectations, but they cannot generate a sustained competitive advantage, especially since the market-demanded applications quickly become commodity products available from numerous suppliers. Lasting ML-based value creation requires creating proprietary systems of record and inquiry, customized ML systems, and platform-based intelligence ecosystems. This article presents strategic recommendations aligned with four distinct ML-based value-creation pathways: table stakes, operational excellence, signature advantage, and stakeholder-value transformation. To support managerial application, we also develop a diagnostic tool to help organizations assess their current data-management maturity, ML-integration capacity, and strategic pathway for ML-based value creation. It stresses the choices necessitated by an organization’s data-management maturity (systems of record) and its existing capacity for ML integration (systems of inquiry). An organization striving to deploy ML applications that elevate its capital productivity should develop a high-value system of record and a well-tailored system of inquiry, which together should position it to generate transformational stakeholder value.