In response to escalating climate-related challenges and evolving regulatory, operational, and environmental demands, metropolitan water and wastewater utilities are increasingly adopting advanced digital technologies. These include smart monitoring systems (e.g., SCADA, GIS, BIM), AI/ML-assisted process control, and twin modeling applications. This paper addresses the growing need for corporate-wide crisis management training for OT system managers and utility operators, spanning executive to technical levels. It presents a methodology for AI/ML-aided risk assessment and scenario-based simulations tailored to organizational structures, vulnerabilities, and operational conditions. The proposed approach supports immersive tabletop training exercises that integrate AI/ML with real-time data and digital twin environments. These simulations enable early event detection, severity rating, and predictive impact assessment across interdependent systems such as hydraulic networks, energy supply, water quality, cybersecurity, and asset management. By leveraging high-frequency operational data, operators can build a dynamic, systemic situation picture to support preemptive decision-making during crises. The methodology includes the use of SCADA-driven BIM platforms with potential GIS/VR integration for time-series visualization of anomalies and operational conditions. Applications are demonstrated for leak localization, bio-contamination detection in water distribution systems, and process control in wastewater treatment. Lessons learned from the design and implementation of these scenario exercises are discussed, offering guidance on integrating innovative AI/ML tools into utility crisis training programs.