Visa has embedded AI into its core infrastructure for more than three decades.

Taking a measured approach to its evolution in payments, the global digital payments leader prioritises security, reliability and continuity, ensuring agent-initiated transactions are built on trusted existing frameworks rather than entirely new systems.

For Mandy Lamb, Head of Value Added Services at Visa Europe, Visa’s long-standing AI foundation is what enables the company to move confidently into the next phase of commerce – one defined by autonomy, intelligence and scale.

Mandy leads Value-Added Services business, which brings together capabilities across fraud prevention, data and analytics, acceptance and payment innovation to help Visa’s clients manage risk and adapt as payments and commerce evolve. 

“Many people know Visa because of cards and payments, but that’s only part of what we do,” she explains.

“At its core, Visa runs the technology that allows money to move securely and reliably around the world. On top of that foundation, we use technology to help financial institutions and businesses change how they operate and grow.”

Through its Value Added Services business, Visa is expanding beyond transactions into transformation.

This includes fraud prevention, data analytics and acceptance solutions that help clients modernise their operations in line with increasingly digital and connected commerce ecosystems.

Mandy adds: “Our work is about supporting transformation – helping clients move from where they are today to where they need to be as commerce and banking become more digital, more connected and more automated.

“As new technologies like AI start to influence how decisions are made and how payments are triggered, Visa deploys tools and experience to help ensure those changes are safe, useful and workable at scale so that merchants can unlock new areas of growth.”

Visa and the agentic commerce push

That transformation is now accelerating with the emergence of agentic AI – systems capable of making decisions and acting on behalf of users.

“As commerce and banking become increasingly digital and automated, technologies like AI are starting to move from supporting transactions to actively shaping them,” Mandy notes.

“We’re now seeing the early signs of agentic commerce, where intelligent systems begin to act on behalf of consumers and businesses.”

Visa’s approach to this shift is grounded in experience.

Its longstanding use of AI in high-stakes environments has shaped a focus on consistency and oversight rather than rapid experimentation.

“Visa has a long history of applying machine learning and advanced analytics at scale, particularly in areas like fraud prevention and risk management,” she explains.

“These systems operate in real time, across billions of transactions, and are designed to balance speed, accuracy and security.

“That experience has shaped a disciplined approach to AI – one that prioritises control, transparency and consistent performance.”

Rather than viewing agentic AI as a departure, Mandy frames it as part of a broader continuum in payments innovation – one where user confidence remains essential.

“In payments, adoption has always depended on trust,” she continues.

“Previous shifts, from e-commerce to contactless, only reached scale once people felt confident that the technology was safe, reliable and working in their interest.

“The same principle applies here.”

Building trust into agentic systems

As systems take on more responsibility in initiating transactions, the underlying infrastructure must provide clear assurances around control and consent.

Payments networks, Mandy suggests, are central to maintaining that stability.

“Payments networks provide the structure that allows transactions to take place safely, with user consent at the core,” she says.

“In an agentic setting, that structure becomes more visible because actions are carried out on behalf of the user.”

Visa’s Intelligent Commerce initiative is designed to extend these principles into AI-driven environments, building on existing rails rather than introducing parallel systems.

“Our framework is grounded by our Visa Intelligent Commerce initiative, where we extend our infrastructure, standards and capabilities present in physical and digital commerce today to AI commerce, enabling AI agents to deliver reliable and secure experiences at every step.”

"New solutions are already being rolled out to support this transition. “We have recently launched Intelligent Commerce Connect to empower businesses to connect and to participate in agentic commerce, supporting them in this next stage of this commerce evolution.”

She emphasises that security remains embedded at every layer of this development, particularly as transactions become more autonomous and less directly user-initiated at the point of action. 

“With this, we help to ensure full security and control via our Trusted Agent Protocol which secures automated transactions within Visa’s network, providing merchants with greater confidence that payments are authenticated and safe.”

She adds that these developments are not being built from scratch, but instead extend the protections already embedded within Visa’s network, ensuring continuity even as capabilities expand.

This includes longstanding security mechanisms that have been adapted to support more autonomous interactions while preserving user intent and oversight. 

“Security measures we utilise, such as tokenisation, are designed to prevent payment credentials from exposure, while authentication and controls help to ensure that transactions reflect the user’s instructions,” Mandy says, noting that these systems continue to operate at scale.

“Visa’s systems validate that agent actions match what has been authorised and apply controls at the network level.

“The same fraud detection and risk systems continue to operate. Agent-initiated transactions are assessed in the same way as any other transaction.”

Testing and scaling agentic innovation

For Visa, the shift towards agentic commerce is being shaped through real-world application rather than theory.

Mandy explains that working directly with partners in live environments has been critical to understanding how these systems behave in practice, particularly when it comes to balancing performance, trust and usability.

Agentic Ready, for example, is a key part of this approach, designed to move beyond concept and into execution while still operating within familiar frameworks.

“One of the clearest lessons is that this needs to be tested in real environments,” she explains.

“Programmes we’ve launched like Agentic Ready are designed to give issuers a structured way to test agent-initiated transactions in live settings, using infrastructure they already trust.

“This helps move the conversation from what might work to what actually does.”
Europe has provided a particularly effective testing ground – not only because of its scale but also due to the complexity and diversity of its payments ecosystem.

This has allowed Visa to trial and refine its approach in conditions that reflect real market dynamics.

“Europe has a unique and dynamic environment that lays a strong foundation for agentic commerce to be tested and grown,” she adds.

Alongside experimentation, Mandy highlights the importance of accessibility, noting that adoption will depend on how easily businesses can engage with these new capabilities.

Reducing integration barriers has been a priority, she says, ensuring that participation does not require significant system overhauls.

She continues: “We have also seen the value of making it easy for partners to get involved. Intelligent Commerce Connect is built as a practical entry point, with a single integration that allows businesses to participate without reworking their existing systems.

“That lowers the barrier and helps the ecosystem move forward at a more consistent pace.”

Looking ahead, Mandy positions agentic commerce as a gradual evolution, where progress is shaped by growing confidence and maturing infrastructure rather than sudden disruption.

Although early signals are already visible, wider adoption will take time as both businesses and consumers follow.

“We don’t see agentic being an all at once change, with developments in this space likely to be layered and gradual,” Mandy shares.

“However, early use cases are already visible, particularly in areas such as product discovery and repeat purchases.”

When it comes to this progression, Visa’s role is focused on enabling scale while maintaining the trust that underpins the payments ecosystem.

For Mandy, innovation must continue to be grounded in the same principles that have historically defined the industry.

“With Intelligent Commerce, we are providing the underlying rails that allow this to scale securely,” she concludes.

“Visa is enabling AI agents to connect to payment systems, authenticate transactions and operate within these defined controls.

Moving forward Visa will continue to work with partners to support the evolution of new ways to pay, ensuring that they continue to rely on the same bases of security and trust – which is what we will always endeavour to provide.”