A TIERED LIFECYCLE GOVERNANCE FRAMEWORK FOR AGENTIC ARTIFICIAL INTELLIGENCE IN FINANCIAL SERVICES

The rapid emergence of agentic artificial intelligence (AI) presents significant governance challenges for financial services, where autonomous systems increasingly support high-impact financial activities. Existing regulatory approaches remain fragmented and continue to rely primarily on non-binding supervisory guidance, leaving important gaps in lifecycle oversight, accountability, and systemic risk management. This paper proposes the Tiered Lifecycle Governance (TLG) Framework, a strategic governance model that integrates lifecycle-based AI oversight with responsive, risk-based regulation. The framework combines three transaction-risk tiers with four AI lifecycle stages to establish proportionate oversight obligations that balance innovation, regulatory accountability, and financial stability. Drawing on technological, economic, legal, and institutional perspectives, the framework is applied to the 2026 U.S. interagency review of model risk management guidance (SR 26-2) to demonstrate how lifecycle governance can strengthen oversight of agentic AI in financial services. The analysis shows that the TLG Framework provides a structured alternative to the existing guidance-versus-rulemaking approach by aligning regulatory obligations with transaction risk rather than institutional size or technology labels. The paper concludes that binding, lifecycle-based oversight offers a practical pathway for improving resilience, transparency, and accountability in financial AI governance while providing a transferable governance model for other emerging AI policy domains.