Agentic commerce—where autonomous AI systems initiate financial transactions on behalf of users within predefined constraints—is moving out of the concept stage and into mainstream payment infrastructure. Visa said on April 29 that its Agentic Ready program is now available in Asia Pacific and Latin America, expanding the company’s efforts to help issuers and partners prepare for agent-led commerce.
What’s significant about Visa’s move is that it centers on the network-readiness problem of agentic payments, rather than framing it as merely an app feature. When a global card network starts treating agentic payments as something issuers need to test and validate, the narrative changes from speculation to real-world application.
Efforts to standardize and regulate agentic payments are moving in parallel. On April 28, the FIDO Alliance announced it was launching its Agentic Authentication Technical Working Group to develop the technical specifications for agent-initiated transactions.
FIDO says its intention is to define how AI agents authenticate, act, and transact on behalf of end users. The problem with existing standards and controls is that they were designed with direct human interaction in mind, rather than the delegated actions of AI agents. Naturally, that has concerning implications for trust, liability, and user consent.
Further pushing agentic payments into the mainstream, major payment-processing provider Stripe also announced at its Sessions 2026 annual conference that businesses can now sell through AI agents via its Agentic Commerce Suite. Stripe Link, the company’s digital wallet and one-click checkout service, can also now support agentic transactions, albeit while still maintaining vital controls like purchase visibility and approvals.
Together, these developments show that agentic commerce is the next big change in the fintech world, as companies shift from pilots to production. Moreover, the infrastructure picture is getting clearer, with agentic payments requiring at least five layers for managing risk throughout the transaction lifecycle. These are: merchant discovery and catalogue access, delegated identity and user intent capture, payment credentialing or tokenization, authorization rules and spend controls, and post-transaction fraud and liability management.
For fintech companies, the market opportunity has never been clearer. That said, there’s much more to the agentic payments race than launching the flashiest new AI-powered shopping assistant. As standards and regulations become clearer, the winners will likely be those who can prove they have the safest autonomous payment rails and the most trusted authorization model.
Just a few months ago, this trend might have looked distant, but it’s fast becoming a product requirement as the major payment infrastructure companies are now leading the way in agentic commerce. For the near-term, the implications are primarily operational, since any fintech that connects to checkouts, wallets, or embedded payments will now need to rethink how they capture user consent.
After all, an AI agent acting under delegated authority is a very different matter from the familiar compliance event of a person clicking ‘buy’. Assuming AI agents ultimately become meaningful traffic sources—as recent developments strongly suggest—fintechs should start thinking now about what trusted delegated payments look like.