trending
Visa is moving to bring agentic commerce closer to commercial reality, launching a new programme with banks across the UK and Europe to test how artificial intelligence agents could initiate and complete card payments on behalf of consumers.
Visa tests foundations of agentic payments
The initiative, branded Visa Agentic Ready, marks one of the clearest signs yet that major payment networks are preparing for a world in which AI does more than recommend products — it also acts on a customer’s behalf at the point of purchase.
The first phase of the programme is centred on issuer readiness.
In practice, that means giving banks direct experience of validating agent-initiated transactions in live, production-grade environments, working in tandem with merchants and Visa’s own infrastructure for tokenisation, identity verification and risk management.
The emphasis is not simply on technical feasibility, but on whether such payments can be executed securely and at scale without weakening the controls that underpin card-based commerce.
Banks are being drawn into early-stage experimentation
Visa has already assembled an extensive roster of participants, including Barclays, HSBC UK, Nationwide Building Society, Revolut and Santander, alongside a broad mix of banks from continental Europe and the eastern Mediterranean.
The breadth of that list matters. Agentic commerce has often been discussed as a future-facing concept driven by technology firms, yet Visa’s approach suggests the financial industry wants banks to remain central to how this model develops.
That is strategically significant. If AI agents are to make purchases on behalf of individuals, issuers will need to determine how consent is captured, how transaction parameters are defined, and where liability sits when automated decisions go wrong.
By involving banks in trials at an early stage, Visa is effectively trying to embed these questions within the existing card ecosystem rather than leaving them to be solved by outside platforms.
Security and consent will determine whether the model scales
Visa’s argument is that agent-initiated payments can be built on rails consumers already know and trust.
The company says tokenisation, biometric authentication and identity tools will help ensure that transactions initiated by AI remain tied to a verified individual and are subject to clear moments of approval and control.
That framing is important because the main obstacle to agentic payments is unlikely to be novelty alone.
Consumers may be willing to let AI search, compare and recommend, but giving software permission to complete a purchase introduces a more sensitive threshold. Trust will depend on whether users feel they remain in command of spending decisions, even when those decisions are partly delegated.
Europe is becoming a proving ground for AI-led commerce
Visa’s decision to begin in Europe reflects the region’s comparatively strong adoption of authentication technologies such as biometrics and passkeys, which are likely to be essential for securing AI-driven transactions.
Early live testing, including a Santander transaction completed by an AI agent using a Visa card, suggests the concept is already moving beyond theory.
For now, the programme remains experimental. Even so, Visa Agentic Ready points to a serious industry effort to shape the rules of AI commerce before it scales.
The question is no longer whether payment networks are preparing for autonomous transactions, but how quickly consumers, banks and merchants will be prepared to follow.
Comments