Start a project

Shopify SEO in the AI Search Era: What Still Matters

4 min read

Đọc bằng tiếng Việt →

Search advice has become noisy. New acronyms for optimizing toward AI answers keep appearing, often with a tool or file you are told to add. For a Shopify store the useful question is narrower: which parts are real, which are marketing, and which old fundamentals still carry the weight? This article sorts the claims using Google’s documentation, Shopify’s developer docs and a recent industry review.

Start with what Shopify fixes for you

Some SEO constraints on Shopify are built in, and it saves time to know them. According to the Fudge review of Shopify SEO in 2026, the URL prefixes for products, collections and pages are fixed. A product can also be reached through a collection path, which is why canonical tags matter; Shopify handles canonicals by default, but it is still worth checking that the duplicate paths resolve properly. Robots rules can be customized through a robots.txt.liquid template. Most themes ship only basic Product and Offer structured data, so richer markup, such as ratings or organization details, is your responsibility. Shopify will not write your collection copy or build your internal links either.

That gives a realistic to-do list: tidy product data, add real content to key collection pages, link to them from articles, and extend structured data where it reflects what is visible on the page.

What Google says about AI features

Google’s documentation on AI Overviews and AI Mode is direct. It says no extra steps are needed to qualify for either. A page needs to be indexed and eligible to show with a snippet, and the standard practices apply: allow crawling, use internal links, offer good page experience, keep important content as text, and make structured data match visible content. For merchants it also recommends keeping Merchant Center information current.

Google’s separate guidance on optimizing for generative AI adds a list of things you can skip. According to it, Google Search does not use llms.txt or similar files, and you do not need to split pages into small chunks, rewrite copy in special phrasing, or add schema just for AI. It also warns against mass-producing pages aimed at every conceivable query, which it says conflicts with its scaled content abuse policy. And it cautions that third-party tools claiming inside knowledge of Google’s ranking systems cannot actually see them.

If you want to control how your content appears, the existing controls apply: nosnippet, data-nosnippet, max-snippet and noindex. Google also says traffic from AI features is included in the Search Console performance report under the Web search type.

Where agents.md fits in

Shopify has introduced a different layer that is easy to confuse with SEO. Its developer documentation describes an agents.md page that tells AI shopping agents how to find and use a store’s commerce capabilities, such as discovery and MCP endpoints, browsing URLs for products, collections and search, and published policies. Shopify manages a default version for every store, and for most stores that is enough. By default, /llms.txt and /llms-full.txt mirror it.

If you need custom instructions, a theme can add an agents.md.liquid template in its templates folder. Two cautions from the documentation: it runs in a restricted Liquid context with only a request object and an agents object, so shop, collection and page objects render blank, and you should never print private merchant details such as contact emails or phone numbers, because the output is broadly cached. A custom file replaces the managed one, and then keeping it current becomes your job.

Note the distinction. Shopify’s file is aimed at shopping agents. Google’s guidance says its Search does not use text files of this kind. Treat agents.md as a commerce feature to review, not as a ranking lever.

A sober look at traffic claims

The Fudge review also cites Pew Research data from 2025: in a sample of US adults, people clicked a traditional result in 8 percent of visits when an AI summary appeared, against 15 percent without one. That is a secondary citation of a 2025 study of US behaviour in one sample, so treat it as context rather than a forecast for your store.

A short priority list for small stores

  1. Complete your product data. Accurate titles, descriptions, variants, availability and identifiers feed Google, product feeds and shopping surfaces alike.
  2. Write for ten collections first. Add genuine buying guidance to your most important collection pages.
  3. Improve internal links from helpful articles to the collections and products they discuss.
  4. Extend structured data carefully. Mark up only what visitors can see.
  5. Watch performance. Heavy JavaScript hurts responsiveness; the Fudge review notes that app-heavy themes often struggle with INP.
  6. Review agents.md once. Confirm the default is acceptable, then leave it alone unless you have a reason.

The Fudge article is published by a vendor and promotes its own product, so we have relied on it for platform facts and the official Google and Shopify pages for the rest. If you want hands-on help with product data, templates or structured data, see our services; for a theme that leaves room for collection content, take a look at our themes.

Further reading

← All posts