The first rewrite of model risk guidance in fifteen years extends supervision squarely to AI. IRIS8 gives banks and credit unions tiered agent controls, defensible inventories, and decision lineage your examiner can follow.
Materiality-tiered model and agent governance, vendor oversight, and documentation templates aligned to the new interagency guidance — with the generative/agentic AI RFI signaling more to come.
AI-aware cybersecurity oversight artifacts, human-oversight evidence, and logging aligned to New York’s expectations for covered entities.
Counterfactual and protected-class testing on decisioning agents, filed as evidence with every certification cycle.
Compiled program procedures grounding AML agents, with escalation rules and full reconstruction for every alert disposition.
Every AI system across lending, servicing, and operations — visible through one gateway and inventoried for your examiner.
Approval queues aligned to your credit authority matrix; action caps and kill switch on every customer-facing agent.
SR 26-2 documentation, adverse-action reconstruction, and examiner packs generated as agents operate.
Bring one policy domain. Leave with a policy set and your first reconstructed decision.
Talk to us