Agentic digital twins: bridging model-based and AI-driven decision-making support for a new era of supply chain and operations management

Agentic AI (artificial intelligence) can profoundly impact model-based decision-making support. This paper conceptualises the notion of agentic supply chain digital twins (A-SCDT) triangulating the composition of agentic AI, digital twins, and model-based optimisation and simulation. Our contribution is twofold. First, we conceptualise the A-SCDT as a distinct and novel area of practical and theoretical importance. Second, we offer a framework named ICARUS (Interaction, Creativity, Adaptation, Reasoning, Ubiquity, and Synchronization), which allows to structure and systematically consider the A-SCDT impacts on future developments of model-based methods in the era of agentic AI systems. Grounding into the ICARUS framework, we elaborate on how the A-SCDT can aid in decision processes describing two industry cases and deducing some generalised insights. We propose a research agenda to stay impactful and relevant in the times of AI-powered supply chains and operations, discussing new topics, barriers, and limitations stemming from AI. Finally, we elaborate on the managerial implications of A-SCDTs and conclude that agentic AI and digital twins are triggering tectonic shifts towards a new era in supply chain and operations management, bridging model-based and model-free, AI-driven decision-making support.