Operational risk management and AI for banks and financial services customers

This article discusses how banks and financial services must adapt their Enterprise Risk Management Frameworks to address operational risks created by generative and agentic AI technologies.

• Generative AI creates different operational risks than traditional ML due to non-deterministic outputs and broader capabilities
• Agentic AI requires operational redesign with proper oversight across functions like operations, risk, and compliance
• Agentic AI risks include conflicting decisions, inconsistent experiences, regulatory breaches, and loss of explainability
• 40% of agentic AI projects may be cancelled by 2027 due to inadequate risk controls and unclear business value
• ERMFs must evolve from static periodic assessments to dynamic continuous monitoring approaches
• AWS provides tools like Bedrock Guardrails, Automated Reasoning, and AgentCore for managing AI operational risks
• Executive priorities include clear leadership, robust infrastructure, cost tracking, and regulatory readiness

Effective operational risk governance enables sustainable AI innovation while maintaining compliance and executive trust in AI deployments.