The agent supports KRI monitoring across the five standard operational risk categories: people risk including staff turnover, training completion, and absenteeism; process risk including error rates, exception volumes, and SLA breaches; systems risk including availability, incident frequency, and recovery times; external event risk including fraud attempt rates and third-party failure events; and compliance risk including regulatory finding rates and policy exception volumes. KRI definitions, thresholds, and escalation paths are configured to match the institution's existing risk appetite framework during deployment.
The agent also supports model risk KRI monitoring, tracking AI model performance, drift indicators, and governance gap flags as a distinct KRI category. Given that the agent itself is a governed model within the institution's portfolio, this category is treated as a first-class risk type rather than an afterthought, consistent with the 2026 interagency model risk management guidance that elevated AI model oversight to an explicit examination concern.
