AI-Driven Root Cause Analysis of Unexpected Operational Deviations: Beyond Error Reduction

While Santoso & Arviansyah (2025) show that AI can help operators adapt to abnormal shutdowns and reduce errors, their work focuses mainly on error mitigation and response optimization. But what if we flipped the script and used AI not just to spot and fix problems, but to deeply analyze why unexpected performance deviations—both good and bad—occur in the first place? This research would design an explainable AI that mines operational process data, not just to prevent mistakes, but to reveal the underlying drivers of outlier events (e.g., why did a team unexpectedly outperform or underperform?). By combining anomaly detection with causal inference and qualitative operator input (inspired by the human-centric lens from Lindner & Reiner, 2023), this research could uncover hidden processes or informal practices that standard metrics miss. Such insights could feed back into training, process redesign, and proactive risk management, making operational learning far more dynamic and nuanced.

References:

1. Strategic Operations under Uncertainty: Mitigating Shutdown Risks and Operator Error through AI-Driven Decision Support. Anton Santoso, Arviansyah Arviansyah (2025). Jurnal Locus Penelitian dan Pengabdian.
2. Industry 5.0 and Operations Management—the Importance of Human Factors. F. Lindner, G. Reiner (2023). IEEE/IFIP Network Operations and Management Symposium.