What is Agentic Commerce?
Agentic Commerce is a model in which an AI agent can take over more stages of the shopping journey,
from finding and comparing products to placing an order and, in some cases, completing payment on the
user's behalf. Instead of searching through individual online stores, a customer can describe what they
are looking for in an AI interface and let the agent research available options, compare them and help
complete the purchase.
The United States is currently the most advanced market for this type of commerce, but Agentic Commerce
is also developing rapidly in the UK and across Europe. In March 2026, Santander and Mastercard
completed Europe's first live end-to-end payment executed by an AI agent in a controlled environment
using Mastercard Agent Pay. The transaction was processed through Santander's live payments
infrastructure and was designed to test how an AI agent can make a payment within predefined
permissions and controls.
This does not mean that fully autonomous shopping is already the standard in the UK. In many current
implementations, AI can search for products, compare offers and build a basket, while the customer still
needs to confirm the transaction before money is transferred. At the same time, payment providers such
as Mastercard and Adyen are developing infrastructure designed specifically for agent-initiated
transactions.
What will Agentic Commerce change in e-commerce?
AI can already act as a shopping adviser. A user can describe what they need, and AI can search for
products, compare available options and identify those that best match the request. This makes the
quality and completeness of product data increasingly important.
An AI agent can base its recommendations on information available about a product, such as price,
features, variants, sizes, materials, dimensions, technical specifications, availability and delivery
conditions. The more precise this information is, the easier it is for an agent to determine whether a
product actually meets the user's requirements.
As Agentic Commerce develops, the role of this data will become even more important. An AI agent may not
only help a customer choose a product but also create a basket, check availability, calculate the final
price and initiate a purchase. The store therefore needs to provide structured and up-to-date
information that can be reliably interpreted by AI systems.
For example, a user might ask an agent to find a lightweight waterproof jacket for autumn, costing no
more than £100, with good breathability and available in size M. The agent can compare products from
several retailers against all of these criteria. If product pages provide clear information about
waterproofing, breathability, weight, materials and available sizes, the agent has the data needed to
evaluate the offers. If the description consists of only a few general sentences, the possibilities for
accurate matching are much more limited.
How will the customer journey change?
In traditional e-commerce, the customer usually enters a store directly, browses products, compares
options and completes the purchase on the retailer's website. With Agentic Commerce, the first point of
contact with a product may instead be an AI interface.
The customer describes what they are looking for, and the agent searches across different retailers,
compares products and presents selected options. This means that a customer can make much of the
purchasing decision before ever visiting a particular store.
The store does not disappear from the process. Its role simply changes. In addition to attracting
customers directly, it needs to be prepared for AI agents to discover, interpret and evaluate its
products.
In some cases, the customer may still be taken to the retailer's website to complete the purchase. In
other implementations, the checkout can take place directly within an AI-driven shopping experience.
Shopify, for example, now provides eligible merchants with Agentic Storefronts that allow products to
appear in AI channels and support purchases through conversational interfaces.
Will SEO still matter?
SEO will continue to matter, but Agentic Commerce may change how we measure its effectiveness. Today,
SEO is primarily associated with increasing a website's visibility in search results and bringing users
to the store. With AI-assisted shopping, a customer can receive a product recommendation without first
visiting the retailer's website.
This means that it will become increasingly important not only for a store to be visible in traditional
search results, but also for its products and information to be discoverable, understandable and usable
by AI systems.
This does not mean that traffic to the website will stop being important. Its role may simply change. A
customer who reaches a store after research performed by AI may already know what they want or have a
much narrower set of products to choose from. Fewer visits can therefore still generate valuable traffic
if those visitors have stronger purchase intent.
The change is already visible in Shopify's data. The company reported that AI-driven traffic to Shopify
stores grew eight times year over year in Q1 2026, while orders from AI-powered searches increased
nearly thirteen times.
What will need to change technically?
Until now, an online store has mainly been built to serve human users. Someone visits the website,
browses products, reads descriptions, adds an item to the basket and pays.
Agentic Commerce introduces another type of user: an AI agent. For the store to work with that agent, it
needs to make its product and commerce data available in a way that AI systems can understand and use.
This can include product information, pricing, stock levels, variants, delivery options, returns, basket
creation, order processing and payment. The exact technical requirements depend on the e-commerce
platform and on the AI or commerce channel the store wants to support.
If you run a Shopify store, much of the infrastructure is already provided by the platform. Shopify has
introduced Agentic Storefronts and Shopify Catalog, which allow eligible merchants to make their
products available across AI shopping channels. Shopify has also introduced the Universal Commerce
Protocol together with Google as an open standard for connecting commerce systems with AI agents.
If you use WooCommerce, PrestaShop or Magento, the situation depends more heavily on the existing setup
and the integration you want to implement. A suitable extension or integration may be enough for some
use cases, while others may require additional API work or a custom connection between the store and
the selected AI commerce platform.
If you run a custom-built online store, the technical work can be more extensive. Developers may need
to create or adapt APIs that allow AI systems to retrieve product information, check availability,
create baskets, calculate orders and communicate with the store's existing order and fulfilment
systems.
Payment integration is also becoming a separate part of this process. It is no longer simply a question
of connecting a standard payment gateway. Providers such as Mastercard and Adyen are developing
infrastructure specifically for agentic transactions, including mechanisms for verifying the agent,
defining its permissions, confirming user intent and securely processing payments. Adyen's Agentic
platform, for example, includes separate layers for product discovery, cart creation and payments.
What about consumer protection and responsibility?
The development of Agentic Commerce does not mean that existing consumer protection rules simply stop
applying. In the UK, the Competition and Markets Authority published specific guidance on the use of AI
agents in March 2026.
The CMA makes clear that businesses remain responsible for complying with consumer law when they use AI
agents. If an AI agent acts unlawfully, the business using it can still be held responsible. This makes
issues such as transparency, accuracy, consumer consent, pricing information, returns and the limits of
an agent's authority particularly important.
For this reason, full autonomous purchasing, where an AI agent completes a transaction without any
human confirmation, is not yet the standard model for UK e-commerce. The technology is developing, but
businesses still need to consider how transactions are authorised, what permissions an agent has and how
consumers remain in control of purchases. Current UK payment infrastructure is being developed around
these requirements rather than simply removing the human from the process.
At the same time, this does not mean that traditional online stores will disappear or that the
human-facing version of an e-commerce website will become irrelevant. Many customers will continue to
browse products themselves, read reviews, look at images, compare specifications and use promotions
available on the retailer's website.
The store will therefore need to support two increasingly connected ways of shopping: the traditional
journey, where the customer interacts directly with the website, and the agentic journey, where an AI
system handles some or potentially most of the research and purchasing process on the customer's
behalf.
