Fix the catalog
Sitemap.xml for AI discovery: helping agents find every product
By Ankit Minocha, founding team at Atomz. Updated January 21, 2026.
A sitemap is how an agent finds the products that are not linked from your homepage. Here is how Shopify's sitemaps work and how to make sure your full catalog is discoverable.
A sitemap is the index that tells a crawler every URL worth reading, including the products buried three clicks deep that nothing on your homepage links to. For a small catalog this rarely matters. For a deep one it matters a great deal, because an agent that only follows on-page links will reach your bestsellers and miss the long tail entirely, and the long tail is often where the specific, high-intent matches live.
A common pattern is a store with a couple of hundred products where only the few dozen linked from collections ever show up in AI recommendations. The rest exist, but nothing pointed an agent at them, and a clean sitemap is how you point.
How Shopify handles it
Shopify generates sitemaps for you automatically. Your root sitemap.xml is an index that links out to sub-sitemaps for products, collections, pages, and blogs, named like sitemap_products_1.xml. You do not have to build them, but you do have to make sure agents can find and use them, which is where most of the value is.
The three things to check are simple. The sitemap should be referenced in your robots.txt so any crawler finds it without guessing. It should be submitted in Google Search Console and Bing Webmaster Tools, since Gemini and Copilot lean on those pipelines. And it should contain your products, which it will unless products are set to a sales channel or visibility that excludes them.
The quiet failure
The most common sitemap problem is not a broken file, it is products excluded from the online store channel, or set to draft, that silently drop out of the sitemap. They look live to you in admin but never reach a crawler. Check channel and status before you check the file.
Keeping it useful
A sitemap helps discovery, but it does not fix readability. An agent that finds a product through your sitemap still has to be able to read and match it, which is the catalog work the other guides cover. Think of the sitemap as the address book and the structured catalog as what happens once the agent knocks. You want both: every product discoverable, and every product worth recommending once discovered.
If you want to see which of your products are both reachable and readable, the audit checks discovery and structure together rather than one in isolation.
Subscribe to The Agentic Operator