<![CDATA[Global supply chains face unprecedented complexity and volatility, accelerating demand for adaptive management systems beyond human-in-the-loop AI. This paper examines agentic AI — AI systems capable of autonomous, goal-directed behavior across multi-step tasks — as applied to end-to-end supply chain orchestration. Unlike conventional AI tools that recommend actions for human execution, agentic AI systems independently perceive supply chain states, reason about optimal interventions, and execute decisions across procurement, inventory, logistics, and risk management functions. Drawing on emerging literature and practitioner experience managing supply chain operations at scale, this paper makes four principal contributions. First, it traces the evolution from assistive AI to autonomous agents, establishing a conceptual foundation for understanding the transformative shift underway. Second, it examines the architectural components of multi-agent supply chain systems, detailing how specialized agents collaborate within enterprise infrastructure. Third, it maps key application domains where agentic AI is generating measurable operational impact. Fourth, it proposes the STAR Framework — comprising Structure, Trust, Adaptation, and Resilience — as a governance architecture for responsible deployment. The paper further introduces an Autonomy Decision Matrix that calibrates agent authority to risk exposure and decision certainty, and critically examines the Deskilling Hypothesis as a counter-theoretical challenge to autonomous supply chain management. Three illustrative case applications validate the framework's practical relevance. Implications for practitioners, organizational designers, and researchers are discussed.]]>
