Strategic Operations under Uncertainty: Mitigating Shutdown Risks and Operator Error through AI-Driven Decision Support

In complex industrial environments such as integrated refinery and petrochemical systems, operational mistakes and unexpected unit shutdowns pose significant strategic challenges. These incidents not only disrupt production continuity but also create substantial financial and safety risks. This study examines the strategic role of artificial intelligence (AI) in enhancing operations management by supporting decision-making during abnormal conditions. A real-case-based simulation was conducted using dynamic models of an integrated plant configuration, with experienced operators performing under both AI-assisted and conventional manual scenarios. Four shutdown scenarios, both detected and undetected, were simulated to evaluate adaptability, multitasking capacity, and resilience. Performance was measured using five key indicators: time to production target, product yield, operational deviations, utility efficiency, and inventory utilization. Statistical analysis employing the Mann–Whitney U test and Bayesian inference revealed that AI assistance significantly enhanced operator performance across all dimensions, particularly by reducing errors and optimizing response times. This research contributes to the ongoing transformation of industrial decision-making by highlighting AI’s potential to improve resilience, reduce cognitive load, and mitigate risks under operational uncertainty. Although limitations related to sample size and scenario diversity are acknowledged, they do not undermine the reliability or practical applicability of the findings for managerial decision support systems.