Both Visa and MasterCard Introduces AI Agents for Shopping: Everything You Need to Know

Artificial intelligence (AI) is no longer a source for finding information. Get ready for artificial intelligence (AI) that not only suggests what to buy but actually buys it for you. 

AI agents are now actively handling our shopping and restaurant suggestions, from browsing to paying, handling the entire booking and payment, all through a chat window. Two of the biggest names in global payment networks, Mastercard and Visa, have announced their latest offerings, AI shopping agents, Mastercard Agent Pay, and Visa Intelligent Commerce. 

Recently, both companies detailed their plans for "agentic commerce," a development that could change how consumers interact with businesses and how transactions happen online. For business professionals, this isn't a distant sci-fi concept; it's a change in commerce that needs close attention.

This is a big step in AI development, where AI no longer just has the role of an assistant but is an active participant in our commercial lives. Both of these payment giants are racing to lay the groundwork for a future where your AI can securely and intelligently make purchases based on users' needs and preferences, ensuring their networks stay at the center of every AI-powered purchase. 

Here's a look at what each agentic AI system brings to the table, how they differ, and what it means for companies and everyday shoppers stepping into "agentic" commerce.

Mastercard Agent Pay

Mastercard recently introduced its Agentic Payments Program, called Mastercard Agent Pay. The idea is to directly incorporate secure payment capabilities into the conversational AI platforms that many already use.

For example, an AI helping someone plan a big birthday bash could suggest outfits, check availability, and then, with approval, buy the chosen items. Mastercard is focusing on making these AI-powered transactions feel natural and trustworthy.

Here's a look at what Mastercard Agent Pay brings to the table:

• Core Function: Integrates AI agents directly into the payment process, aiming for smarter, more personal, and secure transactions for consumers, merchants, and banks.

• Key Technology: Introduces "Mastercard Agentic Tokens." These special digital identifiers build on the company's existing token system used for things like mobile payments to keep AI-initiated transactions safe.

• Partnerships: Collaborating with Microsoft to explore different uses for agent-driven commerce. They're also working with IBM, particularly its Watsonx Orchestrate product, for business-to-business applications, and with payment processors like Braintree and Checkout.com to extend these new token abilities to merchants.

• Security and Control: The program emphasizes that AI agents must be registered and verified. Consumers will maintain control over what their agents can buy, and Mastercard's existing security measures will help to protect against fraud.

• Business Focus: Mastercard wants to keep card-issuing banks central to these new payment methods and to help retailers offer more consistent and personalized shopping by securely identifying customers.

Visa Intelligent Commerce

Close on Mastercard's heels, Visa announced its own initiative, "Visa Intelligent Commerce." Their vision is similar: AI agents that can find products, shop, and complete purchases for consumers.

Visa is offering this as a way to make the entire shopping process, from finding to after-sales support, more personal and secure, and to do it on a large scale.

Visa Intelligent Commerce is built around these key aspects:

• Core Function: Visa is trying to equip AI agents to manage the entire shopping journey for consumers, from finding products and selecting them to making the purchase and handling any post-purchase needs.

• Key Technology: Will provide "Visa Agent APIs" (application programming interfaces – essentially toolkits for developers). These include methods for creating secure digital tokens, authenticating users, and setting spending controls for the AI agents.

• Partnerships: Building a wide network of collaborators, including AI developers like Anthropic, OpenAI, and Mistral AI, tech companies such as Microsoft, IBM, and Samsung, and payment platforms like Stripe. The goal is to create a broad environment for AI commerce.

• Personalization and Support: Highlights the use of many unique signals derived from consumer spending habits to help AI agents make custom recommendations. Visa also plans to offer post-purchase services, like help with disputes.

• Trust and Reach: Focuses on establishing consumer confidence by allowing only designated AI agents to make purchases and by working to set industry best practices. Visa intends to use its extensive global network of cards, merchants, and financial institutions to encourage wide adoption.

How They Work

Both Mastercard and Visa are essentially building the financial rails for AI-powered shopping. They both lean heavily on advanced forms of tokenization—using secure digital stand-ins for your actual card details—to protect transactions.

Consumer control and security are paramount in both proposals, which is vital for building trust. Both agentic AI-based payment systems embed payment credentials into AI-powered conversations, but they take slightly different tacks:

• Mastercard's approach leans on its existing token ecosystem, issuing what it calls "Agentic Tokens." The company lets intelligent assistants request, store, and submit one-time or reusable payment credentials behind the scenes. A party planner could chat with an agent that suggests outfits, then clicks "buy" without handing over card details at checkout.

• Visa's toolkit offers a suite of APIs that let developers bake payment, personalization, and dispute-management services into their AI agents. Instead of a single token product, Visa presents a modular set of tools—authentication, control rules, and even customized recommendations—that agents can mix and match.

Security and Control

Trust is at the core of handing payment power to AI. It means ensuring that business systems can interact smoothly and securely with these new AI intermediaries.

The emphasis on security and clear identification of AI-powered transactions by both Visa and Mastercard should help businesses and consumers navigate this new terrain more confidently. Both Mastercard and Visa have highlighted the following:

• Agent registration: Only verified agents can transact on behalf of users, so you know it's not a rogue chatbot.

• Consumer limits: Shoppers set budgets and item restrictions, from price caps to merchant whitelists.

• Transparent records: Every agent-initiated charge is tagged as such, giving banks and merchants clear visibility.

• Fraud safeguards: On-device biometrics, token revocation, and streamlined dispute flows help stop unauthorized buys and resolve mix-ups.

What It Means for Businesses

This developing field of agentic commerce offers opportunities and considerations for businesses. AI agents could become a powerful and effective new channel for reaching customers, offering highly personalized recommendations and experiences.

• Retailers can offer a richer shopping journey: personalized picks, instant checkout, custom rewards, all delivered by an AI with your trust tokens preloaded.

• Service providers gain an edge by embedding AI assistants that source supplies, negotiate terms, and settle invoices without manual invoices or human error.

• Banks and issuers keep their card products at the core of new commerce models by extending tokenized credentials into untested AI channels.

Conclusion

The announcements from Mastercard and Visa are strong indicators that AI-driven commerce is moving from concept to reality, with AI agents learning to handle more steps, from discovery to delivery, and payments become the final, and most sensitive, hurdle. While these AI agents are still in the early days, the commitment from these payment giants and their tech partners suggests that AI shopping agents will become increasingly common. The key element will be how effectively they can deliver on their promises of security, personalization, and genuine user control. For now, the race to define the future of AI-powered shopping is officially on.