Visa Expands AI Commerce, Stablecoin Settlement and Token Capabilities

Visa has announced new AI, stablecoin and token capabilities at Visa Payments Forum 2026, positioning the launch as part of its strategy to support more automated, programmable and intelligent commerce.

The company said the new capabilities are designed to help clients adapt as artificial intelligence changes the front end of commerce and stablecoins reshape the back end of money movement.

For merchants, banks, fintechs and payment providers, the announcement is relevant because it connects several payment infrastructure trends at once: agentic commerce, tokenized credentials, fraud detection, stablecoin settlement and modular payment modernization.

Visa focuses on AI-driven commerce

A key part of the announcement is Visa’s continued push into AI-driven commerce, where AI agents may help users discover, compare, initiate and eventually complete transactions.

Visa Intelligent Commerce is the company’s platform for agentic commerce. It is designed to provide the trust, controls and connectivity needed for AI agents to securely discover, initiate and complete transactions.

This matters because AI agents introduce new questions into payments. Merchants and issuers need to know whether an agent is legitimate, whether the user authorized the transaction, and whether the payment should be approved without adding unnecessary friction.

Visa’s announcement highlights several capabilities aimed at this challenge, including Agent Score, Agentic Directory and a partnership with OpenAI.

OpenAI partnership brings payments closer to agentic commerce

Visa said it has entered a strategic collaboration with OpenAI to enable secure Visa payments within agentic commerce experiences.

Through the partnership, Visa will provide its global network, credentialing capabilities and security infrastructure to support transactions across OpenAI-powered commerce experiences.

This development is important because AI agents are moving from product discovery toward payment initiation. In traditional e-commerce, the user usually controls every step of the checkout process. In agentic commerce, part of that workflow may be delegated to software.

That creates a new requirement for payment infrastructure: transactions need to carry clearer signals about user intent, agent identity, merchant legitimacy and authorization controls.

For AI products, marketplaces and digital commerce platforms, this could become an important payment layer as more shopping, booking, ordering and service procurement experiences become conversational.

Agent Score and Agentic Directory target trust issues

Visa’s Agent Score is designed to help merchants evaluate whether their websites are ready for agentic commerce. In practical terms, this means assessing whether AI agents can navigate, understand and complete tasks on a merchant’s website.

The Agentic Directory focuses on trust between merchants and agents. Merchants need to understand which agents can be trusted to transact on their sites, while agents need confidence that they are interacting with legitimate merchants.

This is a key infrastructure issue. If AI agents become more active in commerce, merchants may need to distinguish between helpful AI assistants, unknown automated traffic and potentially harmful bots.

For payment teams, this could affect fraud models, authorization flows, checkout design and dispute handling.

Large Transaction Model targets fraud and false declines

Visa also announced a Large Transaction Model, an AI model trained on billions of transactions. The model is intended to improve fraud detection, increase authorization performance and reduce false declines.

False declines are a major issue in digital commerce because they can block legitimate users from completing purchases. At the same time, payment systems must still identify suspicious transactions quickly.

Visa’s Large Transaction Model reflects a broader trend in payments: risk management is becoming more dependent on real-time network data, behavioral signals and machine learning.

For merchants, this matters because authorization performance directly affects revenue. A payment declined incorrectly can create the same business result as a failed checkout: the order is lost, the user experience is damaged, and customer acquisition spend may be wasted.

Token enhancements add more context to digital payments

Visa also announced enhancements to its token capabilities. The company said tokens will carry more data about the transaction type, where the token is being used and who is making the payment.

Visa also introduced a token assurance signal, which evaluates token use throughout its lifecycle based on provisioning and behavioral history.

The goal is to provide issuers with stronger authorization signals. In AI-driven commerce, this becomes more important because the payment may not happen through a familiar human-controlled checkout session.

As commerce extends across apps, devices, agents and embedded channels, payment credentials need to carry more context. Visa’s announcement suggests that identity, permission and behavioral signals are becoming more deeply embedded into payment credentials.

Stablecoin settlement expands across VisaNet

Beyond AI commerce, Visa also shared updates on stablecoin settlement and blockchain-based infrastructure.

The company said it is expanding stablecoin settlement pilots across multiple regions, blockchains and currencies. Building on its earlier stablecoin settlement pilots, Visa said it had moved billions of dollars in stablecoins across VisaNet, with an annualized run rate of approximately $7 billion as of March 2026.

Visa also said issuing banks are already settling seven days a week onchain with Visa, and the company is working to extend seven-day settlement to include acquirers.

This is important for the payment industry because settlement remains one of the most complex parts of global money movement. Traditional settlement rails can be limited by banking hours, currency corridors and operational processes.

Stablecoin settlement does not replace the need for compliance, risk controls or regulated financial institutions. But it can become a useful infrastructure layer for faster and more flexible value transfer between approved participants.

Tokenized deposits point to programmable bank money

Visa also announced work on tokenized deposits, saying it will build a technology layer that can allow banks to turn traditional deposits into programmable, always-on digital money.

This gives banks a way to explore some of the speed and programmability associated with stablecoins while keeping funds on balance sheet.

For financial institutions, this is a significant distinction. Stablecoins and tokenized deposits may both support faster programmable movement of value, but they sit in different regulatory, balance sheet and operating structures.

Visa’s positioning suggests that programmable money is becoming a wider infrastructure topic, not only a crypto or blockchain topic.

Unified Checkout and intelligent authorization support merchant modernization

Visa also emphasized modular modernization for issuers, acquirers and merchants.

For acquirers and merchants, Visa highlighted Unified Checkout, which supports both card and non-card payment acceptance through a single orchestration layer. The company also referenced Visa Intelligent Authorization, which uses real-time network signals and advanced models to help optimize authorization approvals.

This matters because many merchants do not want to replace their entire payment stack at once. Instead, they need modular capabilities that can work with existing systems while supporting new payment models.

As AI-driven transactions, digital wallets, stablecoins and alternative payment methods become more common, merchants may need a more flexible payment orchestration layer that can handle multiple rails and transaction contexts.

Why this matters for digital merchants

Visa’s announcement shows how payment infrastructure is moving beyond basic card acceptance.

For digital merchants, the next stage of payment infrastructure may need to support AI-initiated transactions, tokenized credentials, real-time fraud scoring, programmable settlement and multi-rail checkout.

This is especially relevant for AI platforms, SaaS products, marketplaces, travel platforms, digital services and API-based businesses. These companies often operate across multiple markets, pricing models and user journeys.

In that environment, payment infrastructure needs to answer several questions:

Can the transaction be trusted?

Was it initiated by a verified user, agent or merchant?

Can the payment credential carry enough context for authorization?

Can settlement operate with more flexibility?

Can merchants reduce false declines while still managing fraud?

These questions are becoming more important as commerce becomes more automated and payment flows become more programmable.

The bigger trend: trust becomes the core layer of programmable commerce

Visa’s AI, stablecoin and token announcements all point to the same larger trend: trust is becoming the core layer of programmable commerce.

AI agents can make commerce faster, but they also create new challenges around authorization, identity and control. Stablecoins and tokenized deposits can make settlement more programmable, but they still need compliant infrastructure. Token enhancements can reduce friction, but only if issuers and merchants can interpret the signals correctly.

For the payments industry, the competitive focus is shifting from transaction processing alone to the infrastructure that connects identity, authorization, fraud, settlement and user intent.

Visa’s latest announcement is another sign that the next generation of digital payments will be shaped by how well networks can support trusted, automated and programmable commerce at scale.