Khojasteh Yacob

Recent advances in Agentic AI are beginning to lower several of the principal technical barriers to fully autonomous supply chains. Yet data quality, execution interfaces, formal verification, and governance remain unresolved, and no unified capability framework defines what an autonomous supply chain must accomplish across its full operational and governance cycle, or where current research leaves critical capabilities unaddressed. This paper proposes the SADA-GO (Sense–Analyze–Decide–Act–Govern/Orchestrate) framework, a supply-chain-specific capability taxonomy of sixteen sub-tasks across five domains. It makes two principal theoretical contributions: it elevates Governance and Orchestration to a first-class analytical domain co-equal with the operational stages, and it separates the Analyze and Decide functions that existing frameworks treat as a single function. A structured review of 68 papers published between 2020 and 2026 maps the 44 that provide primary analytical evidence onto the framework's sixteen sub-tasks, while the remaining 24 inform the conceptual background. The analysis reveals an imbalance: autonomous reasoning and decision-making are well researched, whereas the capabilities required to act on those decisions reliably in practice remain largely unaddressed. Across the reviewed empirical implementations, per-decision human approval persists, with no demonstrated transition to human-on-the-loop governance, indicating that cognitive capability and deployable autonomy are not equivalent. Fifteen domain-level and three cross-domain research priorities are derived to close this operational-readiness gap.

Contemporary supply chains operate in a disruption-prone environment where traditional systems, including Supply Chain Control Towers (SCCTs), struggle to ensure resilience. These systems are hindered by two core challenges: 1) data fragmentation and quality issues, which prevent a unified view, and 2) a critical analysis-to-action gap, which delays response by relying on manual intervention. This paper addresses these dual challenges by proposing the "Agentic Supply Chain Control Tower" framework. Using a design science research methodology, this framework features a Digital Twin Tier to mitigate fragmentation and a Cognition Tier containing a workforce of autonomous AI agents to bridge the action gap. The framework's capabilities are validated through a proof-of-concept, demonstrating potential for unified sensing, advanced reasoning, and governed autonomous action. The findings present a new approach for supply chain management based on autonomous orchestration, offering a pathway to more resilient operations.