Organizations are progressively adopting hybrid human–AI systems in decision-making processes in which human decisions are augmented by AI insights. Among the promising AI applications in supply chain monitoring (SCMo) are predictive maintenance systems that predict device failures and augment maintenance decisions, allowing for timely interventions. Despite the growing use of such systems, prescriptive knowledge encompassing technical, business, social, and organizational aspects on how to design, develop, and deploy them in real operational environments remain obscure. To address this shortcoming, our action design research study designed and deployed a predictive maintenance system that predicts the failure of SCMo devices and augments the maintenance decisions of those devices. By doing so, we outline our learnings as generalizable design principles that may guide prospective predictive maintenance systems in SCMo.