Artificial Intelligence as a Governed Capability System for Sustainable Supply‐Chain Management: A Systematic Review, Taxonomy and Research Agenda

Abstract

Artificial intelligence (AI) is increasingly used to forecast demand, evaluate suppliers, optimise production and logistics and support sustainability reporting. Yet much of the literature treats AI as a collection of techniques rather than as a capability embedded in organisational and interorganisational systems. This systematic review synthesises 53 studies at the intersection of AI, supply‐chain management and sustainability‐related outcomes. It separates three analytical products: descriptive findings from the corpus, an abductive theory‐building interpretation and a normative governance extension. The descriptive synthesis identifies a concentration in planning and tactical decision support, with limited attention to delivery, returns, social sustainability, resilience and implemented governance. The theory‐building contribution conceptualises AI‐enabled sustainable supply‐chain management as a governed capability system in which data, models, decision routines, cross‐tier coordination and accountability arrangements jointly shape the conversion of analytical outputs into sustainability outcomes. A process‐mechanism taxonomy links Plan–Source–Make–Deliver–Return activities to sensing, prediction, optimisation and coordination mechanisms. A six‐layer governance architecture is then proposed as a theory‐informed and testable extension rather than as a causal finding of the review. The resulting research agenda focuses on absolute rather than merely relative sustainability gains, evidence provenance, supplier inclusion, labour and distributional effects, resilience trade‐offs and the assurance of generative AI–supported claims.