Case Study: Mastercard and the Rise of Agentic Commerce

Mastercard is pioneering a new era of commerce with the introduction of Agent Pay, an AI-driven payment infrastructure designed for “agentic commerce.” This innovation empowers autonomous AI assistants to execute secure transactions on behalf of users, both in consumer and enterprise contexts. It builds upon Mastercard’s foundational strengths in tokenization, fraud detection, and secure digital payments to deliver a seamless, controlled, and intelligent payment experience. From conversational commerce to enterprise procurement and real-time fraud prevention, Mastercard’s integration of AI technologies reflects a strategic commitment to reshape how individuals and organizations engage in financial transactions.

Key Takeaways

• Agent Pay enables AI agents to make purchases on behalf of users, utilizing secure, tokenized payment credentials.
• Enterprise applications include automated procurement, negotiation, and payment through integrations with platforms like IBM’s watsonx Orchestrate.
• AI is central to Mastercard’s fraud prevention strategy, with tools like Decision Intelligence Pro, Account Intelligence Reissuance, and Scam Protect detecting and blocking fraud in real time.
• Strategic partnerships with Microsoft, IBM, Braintree, and Checkout.com enable Mastercard to embed AI payments across consumer and B2B ecosystems.
• A robust governance and verification framework ensures AI agents act transparently and within user-defined constraints.
• Mastercard is contributing to industry standards for agentic payments to scale the technology responsibly and securely.

Approach

Mastercard’s approach to AI integration is rooted in enhancing its existing infrastructure rather than building standalone solutions from scratch. The company expanded its tokenization framework to create Agentic Tokens, enabling AI agents to operate within trusted transaction environments. This strategic decision allows Mastercard to scale agentic commerce using well-established security mechanisms, such as mobile contactless payments, card-on-file storage, and payment passkeys. Through partnerships with Microsoft and IBM, Mastercard is embedding these capabilities into conversational AI platforms and enterprise software, facilitating frictionless AI-initiated payments in both consumer and B2B contexts. By developing a clear governance model that includes agent verification, user control parameters, and real-time authentication, Mastercard ensures that security and trust remain central to the user experience.

Implementation

The implementation of Agent Pay spans multiple fronts. In the consumer space, Mastercard has enabled AI agents to act as personal shopping assistants — recommending products based on personal preferences, weather, and contextual data, and then securely completing purchases. These transactions are facilitated through conversational AI platforms like Microsoft Copilot, with secure payment credentials managed via Agentic Tokens. In the enterprise realm, Mastercard has collaborated with IBM to integrate Agent Pay into watsonx Orchestrate, enabling AI agents to handle tasks such as negotiating with suppliers, managing cross-border transactions, and automating financial workflows. On the processing side, Mastercard is working with Braintree and Checkout.com to expand existing tokenization systems to support AI-initiated transactions, ensuring compatibility across merchant ecosystems. The company is also building technical standards, such as adapting the Model Context Protocol to Secure Remote Commerce, to differentiate trusted AI agents from malicious actors.

Results

Although Agent Pay is still in its early stages, initial results signal strong potential across multiple domains. In consumer scenarios, AI agents have demonstrated the ability to deliver hyper-personalized shopping experiences while preserving security and user control. For instance, Mastercard showcased how an AI assistant could curate birthday outfits based on individual style and venue details, then autonomously complete the purchase using Mastercard One Credential. In B2B settings, the integration with IBM’s watsonx Orchestrate has opened new pathways for AI-driven automation in procurement, reducing manual effort and streamlining complex financial processes. On the security front, Mastercard’s AI-driven fraud detection system — Decision Intelligence — now analyzes over 160 billion transactions annually, using real-time pattern recognition and risk scoring to block suspicious activities within milliseconds. New offerings like Scam Protect and Account Intelligence Reissuance further bolster Mastercard’s position as a leader in AI-enabled fraud prevention.

Challenges and Barriers

Despite its progress, Mastercard faces several challenges in scaling agentic commerce. The novelty of autonomous AI transactions raises questions about trust, accountability, and regulatory oversight. Ensuring that AI agents are reliably verified, transparent in their behavior, and operate within authorized parameters requires a robust governance framework. Moreover, as AI models rely heavily on historical data, there’s a risk of biased or incorrect decisions — especially in fraud detection — if not properly monitored. Mastercard addresses this through a hybrid model that combines machine learning with human oversight, ethical review boards, and continuous retraining of AI systems. Integrating AI into existing merchant and banking systems also presents compatibility hurdles, which Mastercard mitigates by leveraging its established tokenization infrastructure rather than requiring sweeping system overhauls.

Future Outlook

Looking ahead, Mastercard plans to scale Agent Pay globally, deepen integrations with AI platforms, and shape industry-wide standards for agentic transactions. The company envisions a future where consumers and businesses alike can delegate complex transactional tasks to AI agents with full confidence in security, transparency, and compliance. Mastercard is actively working with partners to establish frameworks that define how AI agents should be authenticated, authorized, and monitored. Innovations like Decision Intelligence Pro, behavioral biometrics, and generative AI-powered fraud detection will continue to evolve, enabling Mastercard to adapt to increasingly sophisticated threats while enhancing customer experiences. As Mastercard’s AI ecosystem matures, it is poised to play a foundational role in shaping the next generation of commerce — one where trust, autonomy, and intelligence coexist seamlessly.

To get the latest AI transformation case studies straight to your inbox, subscribe to AI in Action by AIX — your weekly newsletter dedicated to the exploration of AI adoption in business.

Sources:
Mastercard Launches Agent Pay for AI Payment Transactions
Visa, Mastercard and PayPal Go All in on Agentic Commerce
Visa and Mastercard unveil AI-powered shopping
Visa, Mastercard offer support for AI agents
At Mastercard, AI is helping to power fraud-detection systems
Mastercard Unveils AI-Powered Card Fraud Prevention Service

Let’s talk

Whether you’re looking for expert guidance on AI transformation or want to share your AI knowledge with others, our network is the place for you. Let’s work together to build a brighter future powered by AI.