Summary
FDATA's white paper proposes a governance framework for agentic fintech and write access organized around three categories — Read, Instruct, Transact — each mapped to existing regulatory frameworks covering data protection, liability, and alignment of interests. Rather than proposing a new regulatory regime, the framework applies existing principles proportionately to AI-powered agents that can act on behalf of consumers, including initiating payments.
Key Points
- The core question: Agentic fintech — applications capable of acting on behalf of consumers and small businesses — can expand beyond current read-only data access models to support customer-authorized instructions and payments. The white paper addresses the governance of this write access.
- The 'Read, Instruct, Transact' framework: Agentic activity is organized into three categories: Read (accessing data), Instruct (directing actions), and Transact (initiating payments/moving money). Each category is mapped to existing regulatory frameworks covering data protection, liability, and alignment of interests.
- The pragmatic approach: Rather than proposing a new regulatory regime, the framework applies existing principles. Steve Boms, FDATA Executive Director: "financial regulation does not need to be reinvented for agentic fintech, and that applying existing principles on consumer protection, payments, and data governance in a proportionate and technology-neutral manner can support innovation while maintaining consumer trust and choice."
- The recommendations: Strengthen consumer data rights; enable standing authorizations for appropriately scoped agent-initiated transactions; clarify liability for write-access activities; apply existing duties of loyalty, care, and disclosure in proportion to the level of discretion exercised by AI-powered agents.
- The international context: The analysis draws on international regulatory experience from the UK, EU, Australia, and Brazil to inform policy discussions in the US and Canada, where regulators are developing open finance frameworks that will need to account for AI-enabled financial services.
- The industry collaboration: The report reflects work by FDATA's WAWG, bringing together participants from across the open finance ecosystem — FDATA, MX, Trustly, Sumsub, EQ Bank, Pontera, and Sunset Park Advisory.
- The central governance tension: The paper positions write access — and the governance questions it raises around liability, authorization, and fiduciary duty — as a central issue in the next phase of open finance policymaking. As AI agents become capable of moving money, the balance between innovation and consumer protection becomes critical.
- The significance: The framework provides a template for how regulators can govern agentic fintech without stifling it — applying existing duties proportionately to the level of AI agent discretion. This pragmatic, tech-neutral approach could inform how other jurisdictions regulate AI-enabled financial services.