Agentic Commerce and the Battleground for New Payments Infrastructure

Why It Matters

Agent‑driven payments could dramatically increase transaction speed, volume, and complexity, forcing payment networks and regulators to redesign identity, authentication, and settlement models to maintain trust and stability.

Key Takeaways

• Visa’s four‑step model guides agents from recommendation to full payment orchestration
• New protocols (MCP, A2A, ACP) aim to standardise agent communication and checkout
• Fragmented identity standards risk inconsistency; KYA may replace traditional KYC
• Higher‑frequency, low‑value transactions demand faster, more flexible payment rails
• Regulators must balance AI unpredictability with deterministic settlement requirements

Pulse Analysis

The rise of agentic commerce signals a fundamental re‑engineering of retail payments. As large language models become capable of comparing products, negotiating prices, and authorising purchases, the traditional human‑centric checkout is giving way to autonomous workflows. This shift promises faster, higher‑volume transactions, but also introduces novel risks: agents operate under probabilistic logic, can generate unexpected payment patterns, and blur the line of legal responsibility. Industry players like Visa are already mapping a four‑step adoption path, from recommendation to full‑lifecycle orchestration, to help firms navigate these complexities.

Interoperability is emerging as the linchpin for scalable agentic payments. Current fragmentation—multiple identity frameworks, divergent checkout protocols, and siloed settlement rails—threatens seamless adoption. New standards such as the Model Context Protocol (MCP) and Agent2Agent (A2A) provide a common language for AI agents, while payment‑specific protocols like the Agentic Commerce Protocol (ACP) and Universal Commerce Protocol (UCP) aim to harmonise checkout interactions across card networks, Faster Payments, and blockchain‑based rails. Parallel efforts by Visa Intelligent Commerce and Mastercard’s Agent Pay illustrate the private sector’s push toward unified identity verification, yet the lack of a single governing standard leaves room for inconsistency.

Regulators, particularly the Bank of England, face a dual challenge: ensuring that existing payment infrastructure can safely accommodate agent‑initiated flows and defining new oversight mechanisms where gaps exist. Concepts such as Know‑Your‑Agent (KYA) seek to extend traditional KYC principles to AI actors, while questions around high‑frequency, low‑value transactions demand faster, more resilient rails. Moreover, the deterministic nature of settlement systems must reconcile with the probabilistic outputs of AI, requiring guardrails, policy frameworks, and possibly new central standards. The outcome will shape whether agentic commerce becomes a seamless extension of current payments or a catalyst for a broader overhaul of the financial ecosystem.