Agentic Commerce and UCP and SEO: What Changes for Visibility

For years ecommerce SEO focused on a person typing a query, scanning results and clicking. AI assistants now handle more of that work. A shopper can ask for “running shoes for flat feet under a set budget, delivered by Friday”, and the assistant searches, filters and suggests a short list. With agent checkout features, it may complete the purchase too.

Google has introduced agentic checkout features in its shopping experiences, and other assistants have added shopping and buying features. UCP, announced by Google with commerce and payments partners, is meant to give agents a shared way to talk to merchants, so each merchant does not need separate integrations for every assistant. At the time of writing, support depends on platform and region.

A ranking shows ten results and the user picks. An agent picks for the user, often presenting only a few options. Being sixth used to bring some traffic; being outside an agent’s short list may bring none.

Agents also weigh different things. They compare structured attributes such as price, size, availability, delivery time and return policy directly. Persuasive descriptions matter less; complete and accurate data matters more. Reviews and ratings remain important signals of reliability.

Most of what helps agents is existing SEO and feed work done thoroughly. The difference is that gaps now cost more.

• Product structured data (schema.org Product and Offer) with price, availability, GTIN, brand and reviews
• Merchant listing data, including shipping details and return policy markup
• A complete Google Merchant Center feed, kept in sync with the site
• Clear variant data: sizes, colours and their separate availability
• Policy pages for shipping, returns and warranty that are crawlable and specific
• Fast, stable product pages that render key content in HTML

Feeds become as important as pages

For many agents the feed is the first source, not the page. Errors in Merchant Center, such as price mismatches or missing identifiers, can exclude products. Monitoring feed diagnostics weekly is now part of SEO, not only of paid shopping.

For most merchants, UCP will arrive through their commerce platform, payment provider or Merchant Center integration rather than as custom development. Platforms like Shopify and others have announced support for agent commerce standards; check your platform’s current documentation.

The SEO team’s role is to make sure the data the protocol exposes is correct. If an agent can check out through a standard interface but your product data says “in stock” for items that are sold out, the result is failed orders and lost trust. Inventory and pricing accuracy become visibility factors.

Category and buying guide content still matters, because agents and AI answers draw on it to understand products and needs. Content that clearly states who a product is for, how it compares with alternatives and what its limits are helps agents match products to requests.

• Add specific “best for” and “not suitable for” statements to product pages
• Write comparison pages based on real attributes, not vague claims
• Answer common pre-purchase questions on the page, such as sizing and compatibility
• Keep reviews visible and respond to negative ones with facts

Measurement is still immature. Search Console and Merchant Center reports show performance on Google surfaces, and analytics can show referrals from AI assistants. Orders placed through agent checkout may appear differently in your order system; check how your platform tags them.

Run a regular prompt test: ask several assistants shopping questions in your category and record if your products appear, at what price and with what details. Track errors, such as wrong prices or outdated stock, and trace them to the feed or page.

Agentic buying brings new failure points. An agent may compare you only on price and skip your service advantages. It may misread a variant and order the wrong size. It may pick up an old promotion from a cached page or a third-party listing.

• Keep promotions time-bound in structured data with clear end dates
• Make return and exchange rules easy to find and unambiguous
• Monitor marketplace and comparison site listings for outdated data
• Decide which products you want available for agent checkout and which need a human conversation

Plan for disputes too. Customer service teams should know how to identify orders placed by an agent and what to check first when a customer says the agent bought the wrong item.