Compliance & Standards
Scalable Runtime Governance for Agentic AI in Financial Services
Proposes a scalable governance framework for agentic AI in financial services, arguing that classical model risk management is insufficient because failures emerge during runtime execution. The paper introduces reusable capabilities, pooled validation evidence, telemetry, continuous authorization, policy conformance checks, drift monitoring, and tiered containment.
Topics: payments, regulation, policy, ai, genai