Traditional governance models, originally designed for static software, struggle to manage the shift toward agentic systems that plan, decide, and act with delegated authority. The new initiative addresses this by categorizing AI based on autonomy, reach, and persistence, ensuring controls remain proportional to the potential impact of a system's actions. By moving beyond simple vendor questionnaires, the program offers a bank-led proving ground where institutions can validate AI behavior against specific regulatory and security constraints.

Building on the earlier TrustX Health program, the framework operates across three distinct domains: internal development, third-party procurement, and enterprise protection. Manoj Saxena, founder of the Responsible AI Institute, noted that the industry requires a shared foundation to prove systems operate within approved limits. Participants, including executives from U.S. Bank and NatWest Group, will focus initial efforts on autonomous commerce and payment systems. The project is supported by the TrustX Open AI Registry, an openly licensed resource providing risk classification logic and agent blueprints to standardize governance across the financial sector.