IMF Lays Out How Agentic AI Could Reshape Payments While Preserving Stability

§ 01 Executive Snapshot

• What: The IMF notes the potential of agentic AI to transform payment systems while maintaining stability.
• Who: International Monetary Fund, technology giants, fintech startups, Visa, Mastercard, Stripe.
• Why it matters: This evolution in payment systems introduces significant operational and regulatory challenges, potentially reshaping how transactions are initiated and executed.

§ 02 Key Developments

• A new layer of technical standards like the Universal Commerce Protocol (UCP) and Agent Payments Protocol (AP2) is emerging to support agent-mediated commerce.
• The IMF highlights a three-layer model for agentic AI payments, focusing on intent and orchestration, control and authorisation, and settlement.
• Industry pilots, such as Visa's Intelligent Commerce and Mastercard's Agent Pay, are testing agent-initiated shopping and payment flows.

§ 03 Strategic Context

• Payment processes have historically relied on deterministic logic, ensuring predictability and accountability, while agentic AI introduces probabilistic reasoning.
• The shift from human-initiated to agent-mediated transactions raises new design and policy questions not addressed by existing frameworks.

§ 04 Strategic Implications

• Immediate implications include the need for updated governance structures and technical safeguards to ensure accountability as payments become more automated.
• Long-term, the integration of agentic AI into payment systems may necessitate broader legal concepts of authorisation and trust to manage emerging risks.

§ 05 Risks & Constraints

• A primary risk involves allowing adaptive systems to make irreversible payments without proper controls, checks, or accountability.
• Existing payment frameworks may struggle to adapt to agent-initiated payments, raising questions about traceability, consent, and liability.

§ 06 Watchlist / Forward Signals

• Future developments will depend on the establishment of trusted identity frameworks and interoperable standards for Know-Your-Agent (KYA) verification.
• Policymakers will need to adapt their supervisory approaches to accommodate the evolving landscape of AI-mediated financial activity.