Someone asked Google this question, word for word, and our site showed up for it: "We store important product facts in Shopify metafields. Which AI shopping tools can use those rather than just read descriptions?" It is a good question, and Shopify's documentation answers most of it, though not in one place.
The short answer: Shopify Catalog does not read your custom metafields on its own. It reads your standard product fields. You can bring metafield content in through Catalog's Channel mapping settings, for the title and the description. Everything else in a metafield reaches AI tools only if your storefront shows it, on the page or in structured data, where crawlers can read it.
This post goes through what the Catalog reads, what mapping can change, what Shopify fills in by itself, and what to do with facts that live in metafields. Every statement is from Shopify's docs as we read them on 12 and 13 November 2026.
What Shopify Catalog reads by default
Shopify describes the Catalog as "a structured source of product information gathered from eligible products", used by the Shop app, AI platforms and agents so you do not need a separate feed per channel. The help pages name the fields: titles, descriptions, images, pricing, options and availability. The agentic storefronts privacy page gives nearly the same list for what is shared with the AI channels you turn on: "titles, descriptions, images, pricing, and availability."
Custom metafields are not on either list. Neither are category metafields. Shopify's guide to optimizing products for the Catalog asks for accurate titles, descriptions and images, filled-in product type, vendor, collections and tags, accurate variant option names, a barcode where you have one, and complete store policies. It does not mention metafields.
What Channel mapping can change
Channel mapping is in your admin under Sales channels > Agentic, with Shopify Catalog selected. It is the one place where metafields can feed the Catalog directly. From Shopify's mapping docs:
| Setting | What you can do | What you cannot do |
|---|---|---|
| Product title | Choose the source: "a product field or a metafield". Edit text fields "as Liquid code" and combine sources with your own text | |
| Product description | The same as the title | |
| Product category | "You can't map the product category for Shopify Catalog" | |
| List price | Use the current price, or the compare-at price "only when the compare-at price is higher than the product's current price" | |
| Variant grouping | Group separate products into one listing by title delimiter, by product tag, or by a "Combination metafield" (single line text, product reference list, metaobject reference or metaobject list) | Groups top out at 10,000 products, or 1,024 when grouped by a product reference list metafield |
By default, grouping follows your Shopify Combined Listings setup. A custom grouping "ignores all of your Shopify Combined Listings app setup for Shopify Catalog". Shopify also notes that "any changes that you make are processed with a delay".
The practical use. If your material, dimensions, care instructions or compatibility live in metafields, the description mapping is how you get them into the Catalog. Because the description can combine sources and be edited as Liquid, you can keep your normal description and add the metafield facts after it as plain sentences. That changes only what the Catalog receives, not your product page.
What Shopify fills in by itself
The developer docs for the Global Catalog say the Catalog's responses "can include fields inferred or enriched by Shopify". The listed fields are the description, the normalised options, and a set of metadata: attributes, tech specs, top features, unique selling points, and the condition of each variant. Shopify's own advice to developers is to "treat them as discovery and merchandising signals, not as merchant-authored source text."
Category works the same way. The requirements page says Shopify classifies products using your product information, so the Catalog's category can differ from the one you set in the admin, and you cannot override it through mapping. The Catalog search tool then lets agents filter by category, attributes, rating and ships-to location.
Read together, this is the real answer to the question. Shopify's docs do not list custom metafields among the fields the Catalog reads, so treat a fact that lives only in a metafield as one the Catalog does not have from you. It may still have a version of it, inferred by Shopify from your title, description and images. If you want the Catalog to carry your wording of that fact, map it in. Our post on Shopify Catalog guessing your category shows what an inferred category looks like and how to check it.
Category metafields are a separate thing
Category metafields are the attribute fields Shopify adds when you set a product's category from its standard taxonomy: color, size, material and so on. Shopify's help says they "map to specific product categories" in the taxonomy, that some are added by default "based on AI predictions", and that they help "marketplaces, and search engines". They also drive variant swatches.
Shopify's help does not say the Catalog reads them. It is reasonable to fill them in: they describe your product in Shopify's own vocabulary, and other channels use them. But we have no Shopify source that says filling them changes what the Catalog shows, and we will not claim one.
The other route: your public pages
Shopify Catalog is not the only way AI tools learn about a product. Assistants that browse the web read your public product pages. Shopify's robots.txt help draws the line: Catalog syndication to the agentic storefronts you have turned on happens "independently of /robots.txt", while crawlers reach your store "directly through the open web".
On that route, a metafield is only as visible as your theme makes it. If the theme prints a spec table, a crawler can read the text. If your structured data carries the fact, for example material, color or size on each variant, software reading the markup gets it in a precise form. If the metafield is never rendered, nothing outside your admin can see it.
That is where AgentReady works: on the public, machine-readable facts on your pages, published after you approve them. It does not change your Catalog mapping, and the Catalog's category cannot be mapped at all. The structured data page covers what it publishes.
A short checklist
- List the facts a shopper asks about before buying: material, size, fit, compatibility, care, what is in the box.
- Note where each one lives. Standard field, category metafield, custom metafield, or only in an image.
- For facts only in custom metafields, add them to the Catalog description in Sales channels > Agentic mapping, as plain sentences.
- Make sure the same facts are visible on your product page, and in its structured data where a property exists for them.
- Check eligibility separately. Mapping does nothing for a product the Catalog excludes; our eligibility checklist covers that.
None of this guarantees a product appears in an AI answer. Shopify says each channel "controls the final ranking, wording, and shopping experience." What you control is whether the facts the channel receives are yours and correct. For the wider picture of how the Catalog works, start with the Shopify Catalog guide and the Catalog readiness page.
To see which product facts your public pages already publish, and which are missing, run the free audit.
Comments
Every comment here comes from a verified email. Write yours, confirm from your inbox, and it's live.
Loading comments…