# Search for Mastercard and Google talks to you about agentic commerce, not cards

> Mastercard leaves cards behind on Google and puts agentic commerce front and center as AI payments advance
> By Covadonga Fernandez · Artificial Intelligence · 2026-08-27

A search for Mastercard on Google Spain today, August 27, 2026, presents the company as a global technology firm with tools to power smarter agentic commerce. We did not search for artificial intelligence, agents, blockchain or new payment systems. It is therefore significant that Google places agentic commerce among the first descriptions it offers of a company historically identified with payment cards.

## Mastercard and agentic commerce

Mastercard does not decide the text that appears on Google, since the search engine automatically generates the titles and snippets of its results from various signals coming from the pages and can modify them depending on the query. We therefore do not know the exact combination that led the algorithm to select that content.

What can be verified is that the message matches Mastercard's business strategy. The company has been developing infrastructure specifically intended for agentic commerce since 2025, and Google appears to consider that content relevant enough to show it for a generic search of the brand.

It is worth recalling that the company launched [Mastercard Agent Pay](https://www.mastercard.com/co/es/business/artificial-intelligence/mastercard-agent-pay.html) in April 2025 to integrate payments into a new generation of shopping experiences led by AI agents. The shift from the conventional use of AI is significant, since under previously established instructions and permissions, an agent can represent a person or a company in a transaction.

## Santander and Mastercard

From that capability arises a problem that payment networks have known for decades: trust. If an agent intends to buy on behalf of a person, the merchant needs mechanisms to determine that the agent really represents the user and that it has authorization to carry out the transaction. Spending limits must also be set, credentials protected, transactions authenticated, fraud detected and evidence of consent preserved.

In this sense, Mastercard does not need to build the AI assistant that the consumer will use. Its opportunity lies in providing the infrastructure needed to identify, authorize and process the payment when that agent carries out a transaction. In fact, in March 2026, [Santander and Mastercard](https://observatorioblockchain.com/en/technology/el-plan-de-ana-botin-para-bancarizar-a-los-algoritmos/) announced a complete payment executed in Europe by an artificial intelligence agent using Mastercard Agent Pay and Santander's real payments infrastructure. The transaction was carried out in a controlled environment and allowed the agent to initiate and execute the payment on behalf of the customer within previously established limits and permissions.

## The card loses prominence

In June, Worldline, ING and Mastercard reported an end-to-end agentic transaction in production between an ING customer and a merchant in the Netherlands. The step is relevant because it introduces these technologies into operational payment infrastructures and shows the attempt to move them from testing into real commercial environments.

That same month, Mastercard presented [Agent Pay for Machines,](https://www.mastercard.com/us/en/business/artificial-intelligence/mastercard-agent-pay/agent-pay-for-machines.html) an infrastructure to facilitate payments made between machines and autonomous agents. The company envisions scenarios in which these systems can contract services or execute programmatic transactions, including numerous small-value operations carried out at machine speed. It is an emerging technology, but it shows how far Mastercard intends to extend the concept of machine-to-machine payment.

From here, it could be said that the card is losing prominence as a definition of Mastercard. The reason is that it is no longer enough to describe all the activities with which the company seeks to position itself in the new payment systems. First it dominated the physical card, then its credentials were incorporated into phones and digital wallets.

## Comparison with Visa

Tokenization made it possible to protect and use those credentials in new ways. Now Mastercard wants them to be integrated into operations initiated or managed by authorized AI agents. What changes is the invisible infrastructure behind the act of paying.

The comparison with its competitor Visa helps confirm that the phenomenon goes beyond Mastercard. When searching for Visa on Google Spain, the main result also does not use cards as the central concept to present the company. Visa appears as a trusted leader in digital payments and the description talks about a digital and mobile payment network and new payment technologies.

Although it does not appear in Google results at first glance, Visa is also moving fully into agentic commerce. The company is developing Visa Intelligent Commerce and technologies aimed at identifying, authorizing and protecting operations initiated by AI agents. In June 2026 it announced a collaboration with OpenAI to incorporate its payment network, credentials and security capabilities into agentic commerce experiences.

## The shift that AI poses

Google's results do not allow the conclusion that Mastercard bets on agents while Visa bets on trust. Both companies work on both fronts. The difference lies in how their main results appear for a generic Google search. Mastercard is explicitly associated with agentic commerce and Visa with trust and digital payments, but in neither case does the card occupy the center of the presentation.

The shift that AI poses may be greater than the one caused by the partial replacement of plastic by the mobile phone. E-commerce changed where we shopped. The smartphone changed the device used to do it. Agentic commerce may change who executes certain phases of the purchase.

A consumer can ask an agent to find a hotel within a given budget, compare conditions and select an alternative. If it has authorization to continue, that same system can move forward to booking and payment within the established limits.

## Stablecoins and the crypto industry

In the business and user sphere, specialized agents are already beginning to contract digital services and execute payments to complete tasks. And here another battle opens up. Santander and Mastercard showed that an agent can pay using traditional financial networks, without blockchain or stablecoins. The crypto industry is developing a parallel path. For example, Coinbase allows agents to operate with wallets and pay for services using stablecoins through x402, while Circle has created infrastructure for programmatic payments with USDC. The question now is which rails the money of AI agents will travel on.
