Operationalizing Responsible AI in Regulated Banking: A Persona-Driven Control Framework

The adoption of Artificial Intelligence (AI) in regulated banking environments has rapidly expanded beyond traditional use cases into fraud detection, anti-money laundering (AML), cybersecurity analytics, credit decisioning, and generative AI–driven customer interaction systems. While AI enhances operational efficiency and predictive capabilities, the use of AI introduces systemic risks in highly regulated financial institutions because AI models and their resulting decisions must be explainable, auditable, fair, and defensible under supervisory review. There are currently many existing responsible AI frameworks that emphasize governance principles such as transparency, fairness, and human oversight but often lack implementation-level clarity across enterprise roles. This paper proposes a persona-driven responsible AI control framework that distributes accountability across enterprise stakeholders including developers, platform engineers, security teams, operations personnel, and business decision-makers. The framework embeds enforceable control planes within enterprise architecture, transforming governance from policy documentation into measurable execution. By integrating data governance, model lifecycle controls, identity enforcement, runtime observability, and operational resilience mechanisms, the framework bridges governance policy with engineering implementation in multi-account regulated cloud environments. The approach provides financial institutions with a structured, auditable pathway for scalable and compliant AI adoption.