Worldline, ING and Visa say they completed a live “agentic” payment transaction in Europe

Worldline, ING and Visa announced that they have completed a live agentic payment transaction in Europe, positioning the demonstration as a proof point for “agentic payments,” payments initiated with the help of autonomous AI systems rather than only traditional, step-by-step user interfaces. The firms said the use case was designed to operate securely with strong customer and biometric authentication, according to the announcement carried by Yahoo Finance.

In practical terms, an “agentic” payment effort is aimed at shifting some of the work of initiating a purchase or payment from rigid workflows to an AI agent that can carry out parts of the transaction process, such as interpreting intent and preparing an authorization request. Supporters argue this could reduce friction in payments while maintaining compliance controls, but it also raises new security and auditability questions that current payment networks and banks must address.

Worldline’s materials, as summarized in the market coverage, framed the demonstration around end-to-end execution in a live setting and emphasized secure and compliant operation. The companies said the transaction illustrated how agentic payments can work using strong customer and biometric authentication, a reference to mechanisms that tie the transaction to a verified customer identity, and in this case include biometric checks.

ING is a major European bank and a participant in payments infrastructure and digital customer journeys. Visa, as a global payments network, plays a role in how transaction messages and authentication results are coordinated across systems. Worldline, a payments and financial services technology provider, typically supplies acquiring, transaction processing and related software, and in this case is presenting the work as something that can be deployed rather than a purely laboratory concept.

The announcement also highlights that the transaction was “live” and carried out in Europe, suggesting the firms were not only testing AI decisioning, but also connecting it to authentication and payment execution processes that mirror operational systems. The Yahoo Finance version described the event as a new use case illustrating secure operation with strong customer and biometric authentication, and other coverage referenced similar “end-to-end” language around a European production environment.

The companies did not, in the available public excerpts, specify the exact merchant scenario, the payment rails used, the countries involved, the timeline, or performance outcomes such as authorization rates or processing latency. They also did not disclose what portion of the transaction flow the AI agent handled versus what remained behind conventional bank-verified customer steps.

For now, the best-supported takeaway is that the demonstration is meant to address two persistent blockers for AI-assisted payments: authentication strength and secure handling of transaction intent. By explicitly pairing agentic execution with biometric and customer verification, the firms are indicating that they see authentication as a central control, not an afterthought. That is consistent with broader industry concerns that AI agents must be constrained, auditable, and tightly integrated with established security processes.

What to watch next is whether the companies publish more detail about how the AI agent’s actions are bounded and audited, whether there are regulatory or certification steps tied to the approach, and whether additional banks or merchants adopt the workflow beyond pilots. Buyers and partners in this space will likely look for clearer documentation on safeguards, error handling, and how customer consent is recorded when an agent initiates or completes transactions. The longer-term question is whether “agentic” payments move from demonstrations to repeatable deployments that satisfy both security expectations and operational requirements.