Santander and Mastercard Complete Europe's First Live AI Agent Payment
Banco Santander and Mastercard (NYSE: MA) announced on March 2, 2026, the successful completion of Europe's first live end to end payment executed entirely by an artificial intelligence agent. The transaction, processed through Santander's live payments infrastructure in Madrid and London, also represents the world's first agentic payment carried out within a regulated banking framework [1]. The milestone was achieved using Mastercard's Agent Pay solution, orchestrated by PayOS , Santander's payments platform. Unlike sandbox demonstrations or simulated environments, the AI agent initiated and completed a real payment on behalf of a customer without human intervention at execution time, operating within predefined spending limits and permissions set by the customer and the bank [2]. How the Transaction Worked The payment flow began with an AI agent registered with identity credentials, spending limits, and a full audit trail. The agent operated as a visible, governed participant in the payment chain, interacting with issuers, acquirers, and merchants through Mastercard's existing tokenization infrastructure. Cryptographic proof of consumer intent was generated for every party in the transaction, from the merchant to the issuer. PayOS provided the end to end orchestration layer, while Mastercard Agent Pay ensured the AI agent could be recognized and identified at every step [1][2]. "At Santander, we see AI as a transformative force in the evolution of payments. Our role is not only to adopt innovation, but to shape it responsibly, embedding security, governance and customer protection by design. As AI agents become part of everyday commerce, building trusted, scalable frameworks will be essential to unlocking their full potential." [1] Matías Sánchez , Global Head of Cards and Digital Solutions at Santander, made those remarks in the joint announcement. His counterpart at Mastercard, Kelly Devine , President for Europe, emphasized the continuity between traditional payment principles and the new agentic paradigm. "Agentic payments represent a profound shift in how commerce is initiated and executed. With Mastercard Agent Pay, we are applying the same principles that have defined our network for decades, security, trust, interoperability and global scale, to a new era of AI enabled commerce." [2] Mastercard Agent Pay: Architecture and Mechanics Agent Pay was first unveiled on April 29, 2025, in Purchase, New York, and integrates with Microsoft Azure OpenAI Service and Microsoft Copilot Studio . The system builds on Mastercard's existing tokenization stack, the same infrastructure behind contactless payments, card on file transactions, and Payment Passkeys [3]. | Component | How It Works | | | | | Registration and Authentication | AI agents are registered and verified before making any payment | | Agentic Tokens | Built on Mastercard's existing tokenization infrastructure | | Consumer Control | Users set predefined limits on what AI agents can purchase | | Transparency | Every participant (consumer, issuer, merchant) can identify agentic transactions | | Fraud Protection | On device biometrics used for strong consumer authentication | | Agent ID | Issuers can reject transactions lacking agent identification | | Cryptographic Proof | Consumer intent is cryptographically verified for all parties | The controlled pilot at Santander did not represent a commercial rollout. The bank said it would move into "extended testing and scaling" and explore additional use cases and partnerships, though no timeline for public availability was disclosed [1]. Global Context: Banks Testing Agentic Payments Santander's achievement arrives amid a broader wave of financial institutions trialing Mastercard Agent Pay. Citi and US Bank are both conducting trials of the technology in the United States, while Westpac in New Zealand completed that country's first authenticated agentic transaction in approximately February 2026 [3][4]. | Bank | Market | Status | | | | | | Banco Santander | Europe (Spain/UK) | First live regulated AI agent payment (March 2, 2026) | | Westpac | New Zealand | First authenticated agentic transaction (February 2026) | | Citi | United States | Trialing Mastercard Agent Pay | | US Bank | United States | Trialing Mastercard Agent Pay | Analysts at eMarketer described the Santander pilot on March 4 as creating a "new power dynamic in payments: protocol dominance," noting that agentic commerce could shift the competitive landscape toward infrastructure providers who establish the rules for machine to machine transactions [5]. Santander's Scale and AI Ambitions Banco Santander , founded in 1857 and headquartered in Spain, operates across markets with 178 million customers , 7,400 branches , and 201,000 employees . The bank holds approximately EUR 1.3 trillion in total assets as of Q3 2025. Its AI strategy is led by Ricardo Martín Manjón , who joined as Chief Data and AI Officer in March 2025 after leaving BBVA [1][6]. FF News described the pilot as "far more than a technological proof of concept," calling it "the crucial validation that existing, regulated payments infrastructure can safely accommodate autonomous, machine driven commerce" [4]. That distinction matters because agentic payments introduce novel risks around authorization, liability, and fraud that regulators have not yet fully addressed. By executing within a live, regulated environment, Santander and Mastercard have demonstrated that existing compliance frameworks can accommodate AI driven transactions without requiring entirely new infrastructure. Mastercard's broader AI partner ecosystem includes IBM (watsonx Orchestrate), Braintree , and Checkout.com , in addition to Microsoft. The company's Chief Product Officer, Jorn Lambert , has described Agent Pay as designed for an era when billions of AI agents will transact on behalf of consumers, making identity, trust, and interoperability the foundational challenges of the next decade …