Santander and Mastercard Launch the First AI-Powered Payment Pilot in Europe

Artificial intelligence rapidly transforms every industry including the financial services sector. Banks are using AI to optimize countless activities including customer service enhancement, trading, and fraud detection. New advancements include the ability of AI to autonomously process financial transactions.

For the first time in Europe, an AI agent has completed an end-to-end payment within the regulated banking system. This major milestone, as part of a pilot project, marks the initiation of a new dimension of agentic commerce, where AI can autonomously process a payment from start to finish.

The experiment illustrates the potential of autonomous software to conduct routine purchases, subscriptions, and business transactions while operating within strict security and regulatory frameworks.

An Unprecedented First in AI-Powered Payments

The pilot transaction took place over Santander’s live payment infrastructure using Mastercard’s Agent Pay platform. This setup allowed the companies to determine whether AI could successfully process a payment in a real banking environment.

The goal of the test was to verify whether the entire payment process could function correctly, starting from initiation to authorization and finally completion. Most importantly, the payment had to meet all operational, security, and compliance requirements within the regulated environment of Santander.

The successful completion of the transaction demonstrated that AI agents can operate within the existing financial system while maintaining the safeguards necessary to protect consumers.

Many observers view this pilot as one of the first significant examples of autonomous payments within the European banking system.

What Are AI-Executed Payments?

AI-executed payments involve autonomous software agents performing transactions on behalf of users. The account holder defines the rules, permissions, and spending limits that the AI must follow.

Instead of requiring the user to manually approve each transaction, the AI agent operates automatically once it has been configured according to the user’s preferences.

Examples of actions an AI agent could perform include:

• Automatically reordering household supplies when they run low

• Renewing digital subscriptions

• Monitoring corporate expenses

• Paying service bills or invoices based on agreed payment terms
  

In this model, AI systems effectively function as users within the payment ecosystem, interacting with banks, merchants, and payment networks.

However, these systems are not completely unrestricted. Transparency and control remain essential. Customers can define spending limits, establish rules, and monitor the activities carried out by their AI agents.

Operation of Mastercard’s Agent Pay Platform

Mastercard’s Agent Pay platform is designed to integrate AI agents into the payment process as authenticated participants.

Traditionally, payment networks connect several entities:

• Merchants and cardholders

• Issuing banks

• Acquiring banks

• Payment processors
  

Agent Pay introduces a new participant into this ecosystem: the AI agent.

The platform ensures that AI agents are properly identified and authenticated within the payment network. Instead of directly accessing payment card details, the system uses Secure Payment Tokens (SPTs), which represent the card without exposing sensitive information.

This tokenization approach allows the AI agent to execute transactions while protecting privacy and reducing the risk of fraud.

During the pilot program, PayOS coordinated the transaction between the banks, merchants, and payment networks involved.

The Importance of This Innovation

The successful pilot represents more than just a technical achievement. It signals a shift in how digital commerce may function in the future.

Increased Autonomous Commerce

Future payment systems may allow AI-powered devices to automatically purchase goods and services.

For example, a business system could automatically manage inventory purchases when stock levels fall below a certain threshold. Similarly, smart home devices such as refrigerators could automatically reorder groceries.

These types of scenarios demonstrate how autonomous payments could simplify purchasing processes and improve efficiency.

Greater Efficiency for Businesses

Businesses could benefit from automating recurring payments and financial processes such as subscription management and supply purchases.

This automation reduces the need for manual administrative work and streamlines financial operations, saving both time and operational costs.

Greater Convenience for Consumers

Routine financial tasks could be handled automatically by AI assistants, reducing the need for manual involvement.

For instance, an AI assistant might monitor airline ticket prices and automatically book travel when prices drop below a predetermined threshold.

By handling repetitive tasks, AI systems could make everyday financial management significantly easier for consumers.

Governance and Security Remain Critical

Despite the advantages, autonomous payments raise important security and legal considerations.

AI systems operating within the financial sector must be carefully monitored to ensure they cannot bypass authentication systems or misuse customer funds.

The Santander pilot addressed these concerns by allowing AI agents to operate only within tightly controlled regulatory and security frameworks. The AI agent was given limited flexibility and could perform transactions only according to predefined guidelines.

This ensured compliance with strict safety standards and consumer protection requirements.

European regulations such as the Payment Services Directive 2 (PSD2) require Strong Customer Authentication (SCA) for electronic payments. As a result, AI-driven payment systems must also comply with advanced identity verification procedures.

The pilot demonstrated that autonomous payment systems can function within these regulatory constraints while maintaining high levels of security.

The Future of AI in Financial Transactions

Although the pilot represents a major milestone, the technology remains in its early stages.

Following the successful test, Santander plans to continue developing AI-driven transaction models through additional trials, new use cases, and industry collaborations.

Meanwhile, financial institutions and payment networks worldwide are exploring similar innovations.

The concept of agentic commerce—where AI systems autonomously conduct transactions—is gaining attention across the global financial industry.

Other ongoing developments include:

• AI-powered fraud detection systems

• Autonomous algorithmic trading platforms

• AI-driven financial assistants

• Machine-to-machine payments

• Internet of Things (IoT) devices capable of conducting transactions
  

Together, these innovations are accelerating the automation of financial services.

AI-Driven Commerce in a New Era

The collaboration between Santander and Mastercard highlights how artificial intelligence is rapidly moving from analytical support roles into the operational core of financial systems.

For years, AI has helped banks analyze data, detect fraud, and improve risk management.

Now, AI systems are beginning to execute transactions themselves.

With responsible development, strong governance, and secure technological frameworks, AI agents could eventually become active participants in the global economy.

In the future, AI may not only recommend products and analyze spending patterns but also manage financial transactions on behalf of users.

The successful European pilot suggests that this new era of AI-driven commerce may arrive sooner than expected.