Mastercard and Agentic Commerce: When AI Starts Shopping for You

AI Is Moving From Recommendations to Transactions

Artificial intelligence has already changed the way consumers discover products.

People increasingly ask AI assistants to compare hotels, recommend software, research electronics or find the best product for a particular budget. Until recently, however, the final steps of the journey still belonged largely to the consumer: visit the website, add the item to the cart and complete payment.

That boundary is beginning to disappear.

Mastercard is expanding its infrastructure for agentic commerce, a model in which AI agents can perform parts of the shopping process on behalf of a person — including product discovery, comparison and, with authorization, completing a transaction.

The difference is significant.

An AI assistant can tell you which laptop might be right for you.

An AI commerce agent could potentially find that laptop, check availability, compare offers, select an approved merchant and pay for it within the rules you previously established.

That turns AI from a recommendation engine into an active participant in commerce.

Mastercard Pushes Deeper Into AI-Powered Payments

On September 17, The Wall Street Journal reported that Mastercard is working with Alchemy on technology that can give AI agents virtual payment credentials.

Instead of providing an AI system with unrestricted access to a conventional credit card, consumers can establish conditions around what the agent is allowed to do — including spending limits and restrictions on where purchases can be made.

The idea is essentially to give an AI agent purchasing authority without giving it unlimited financial authority.

This fits into Mastercard's broader strategy around Mastercard Agent Pay, its infrastructure designed specifically for payments initiated through AI-powered agents.

Mastercard says the system is intended to combine AI-driven commerce with authentication, consumer authorization and controls that establish what an agent has permission to purchase.

The Travel Industry Is Already Testing the Model

Travel provides an easy example of what agentic commerce may look like.

On September 16, 2026, Mastercard and Trip.com announced a collaboration with Network International showcasing an AI-powered travel purchasing experience.

Trip.com's AI agent, TripGenie, is being used to demonstrate a journey in which consumers can search for travel services, evaluate options, book them and complete purchases through an AI-powered experience.

Importantly, this should not be confused with a universal consumer service that is already fully deployed. Mastercard says the companies plan a phased rollout, with commercial launch expected in early 2027.

But the direction is clear.

A future travel request could look something like:

"Find me a hotel in Tokyo near Shinjuku Station for five nights, rated at least 4 stars, with breakfast included, for less than $1,200."

Instead of returning ten links and leaving the rest of the work to the traveler, an AI agent could eventually research the options, confirm the conditions and complete the reservation — as long as the transaction fits the permissions established by the customer.

The Bigger Story Is Not Payments. It Is Product Discovery.

For marketers and eCommerce businesses, the most important part of this development may not actually be the payment technology.

It is the possibility that the buyer evaluating a product may increasingly be software.

On September 9, Mastercard announced an expansion of its Agent Suite for Merchants and introduced Mastercard Agent Connect.

The company specifically describes helping merchants make their products discoverable across AI ecosystems.

Participating AI agents can access merchant-provided information such as product details and availability, recommend products, confirm prices, taxes, shipping and fulfillment information, and ultimately support transactions.

That introduces a completely new marketing question:

How does a business convince an AI agent to select its product?

For decades, digital marketing has largely been built around influencing human behavior.

Search ads were designed to attract clicks.

Landing pages were designed to persuade visitors.

Product photography was created to attract attention.

Reviews built consumer confidence.

SEO helped humans discover businesses through search engines.

Agentic commerce introduces another audience into that system: machines making decisions on behalf of humans.

From SEO to AI Commerce Optimization

This does not mean traditional websites, branding or advertising are suddenly irrelevant.

It means another layer may be added to the customer journey.

Businesses could increasingly need to ensure that their products are understandable not only to people, but also to AI systems.

That could make several areas increasingly important:

Structured product data
AI agents need accurate information about products, specifications, variants and categories.

Current pricing and availability
An agent making a purchasing decision needs reliable real-time information rather than outdated product pages.

Shipping and fulfillment information
Delivery dates, shipping costs and geographic restrictions can become part of an automated purchasing decision.

Return and refund policies
Clear policies provide another machine-readable factor that an agent may evaluate when comparing sellers.

Merchant reputation and trust signals
Trust becomes particularly important when customers delegate purchasing decisions to software.

Product feeds and integrations
Businesses may increasingly need their catalogs available through systems that AI platforms and commerce agents can reliably access.

Accurate Schema and structured metadata
Machine-readable data can help AI systems understand what a business sells and the attributes attached to each product.

The traditional objective of SEO has been:

Help the search engine understand your content so it can show it to a person.

An emerging objective of agentic commerce may become:

Help the AI understand your product well enough to select it for the person.

What Happens to the Customer Journey?

The classic digital customer journey often looks something like this:

Advertisement → Search → Website → Product Page → Cart → Checkout

Agentic commerce could compress much of that journey:

Request → AI Agent → Product Selection → Authorized Purchase

In some cases, the customer might never browse the merchant's website in the traditional sense.

That would have major implications for attribution, analytics, advertising, conversion optimization and even the meaning of website traffic.

A company could theoretically generate revenue from customers who never personally viewed its product page.

That presents both an opportunity and a challenge for marketers.

The opportunity is access to entirely new AI-driven distribution channels.

The challenge is maintaining brand identity, customer relationships and marketing attribution when an intelligent intermediary stands between the company and the consumer.

Mastercard appears aware of that tension. Its Agent Connect strategy emphasizes allowing merchants to retain control over pricing, fulfillment, customer relationships, brand experiences and the AI environments in which they participate.

Trust Could Determine How Quickly Agentic Commerce Grows

Giving software permission to spend money introduces obvious questions.

What happens when an agent purchases the wrong item?

Who is responsible if the agent misunderstands a request?

How are fraudulent AI transactions identified?

How can merchants distinguish a legitimate authorized agent from malicious automated activity?

And how much purchasing authority will consumers actually be comfortable delegating?

These questions are one reason payment companies are focusing so heavily on authorization and identity.

Mastercard's approach includes technologies intended to verify consumer intent and establish clear boundaries around what an AI agent has permission to do.

The technology may advance quickly, but consumer trust will ultimately determine how much purchasing authority people are willing to hand over.

Marketing Is Entering the Agentic Era

The rise of AI search already forced marketers to think beyond conventional Google rankings.

Agentic commerce could represent the next stage of that transition.

The question may no longer be only:

"Can customers find our company?"

Or even:

"Will ChatGPT, Gemini or another AI platform mention our company?"

It may increasingly become:

"Will an AI purchasing agent choose our company?"

That is a fundamentally different problem.

It combines SEO, eCommerce, product data, reputation, payments, APIs and artificial intelligence into a new type of digital customer journey.

Mastercard's latest moves do not mean human shopping is disappearing.

They do suggest that a new participant is entering the marketplace.

And the next customer a business needs to optimize for may not be a person visiting a website.

It may be an AI agent acting on that person's behalf.

---

Sources

The Wall Street Journal — Mastercard Joins Visa in Letting AI Bots Do Your Shopping
Read the WSJ report

Mastercard — Mastercard and Trip.com pave the way for agentic commerce in the travel sector
Read the Mastercard announcement

Mastercard — Agent Connect and Agent Suite for Merchants
Read the Mastercard merchant announcement

Mastercard — The Future of Shopping and Payments
Read Mastercard's agentic commerce report