Merchant field guide · 01

How to check your Shopify store for agentic shopping

Start with the product source shopping agents can actually query. Then improve the catalog facts they receive, publish the policies buyers need, and verify the public web layer. A special text file cannot make up for weak product data.

Reviewed July 25, 20269-minute readNo ranking guarantees

The short version

Readiness is a chain, not a switch.

For Shopify's agentic storefronts, Shopify Catalog is the primary product-data route. It can provide titles, descriptions, options, images, prices, availability, and other attributes in a structured form. Open-web crawling is a separate route. Your storefront, sitemap, robots rules, and product markup still matter there, but they do not replace Catalog access.

Treat the layers below as a dependency order. A healthy public page is useful evidence; it is not proof that a product is eligible, indexed, selected, ranked, or recommended by any channel.

Layer 1

Confirm Shopify Catalog access and channel eligibility

In Shopify Admin, open Sales channels → Agentic and review the channels that can access your products. The default managed setting can enroll eligible stores in available agentic storefronts; merchants who turn it off can manage Catalog access and supported checkout options more granularly. Availability and purchase flow differ by channel, region, store, and product.

What to verify

  • The intended AI channel has Shopify Catalog access, and the store meets that channel's current eligibility terms.
  • Products are active, available to the relevant market, and not deliberately hidden from discovery.
  • A realistic query in Shopify's Catalog search preview returns the products you would expect.

Shopify describes its search preview as directional. Agentic channels can re-rank results using their own logic, so a preview position is not a promise about what a shopper will see.

Layer 2

Make product facts complete enough to answer a buyer

Shopping queries are often specific: material, fit, use case, compatibility, size, color, certification, budget, or delivery constraint. Put decision-making facts in the product fields Shopify can syndicate. Do not rely on a lifestyle image or an unexplained variant code to carry the meaning.

Identity and meaning

Use a precise title, an original description, the most accurate product category, and relevant attributes. Explain what the item is, who it suits, and important limitations.

Images and variants

Give important variants understandable option names, current availability, and useful imagery. Alt text should describe the image, not repeat keywords.

Commercial facts

Keep price, currency, inventory, SKU or product identifier, and offer availability consistent. Stale commercial data undermines an otherwise detailed listing.

Custom data sources

If essential titles, categories, or attributes live in metafields, metaobjects, tags, or custom grouping logic, review Shopify Catalog Mapping instead of assuming the default source is correct.

Shopify's current listing-quality indicators call out description completeness, image coverage, verified reviews, variant and option completeness, and shop-policy completeness. These are controllable inputs—not a formula that guarantees a position.

Layer 3

Publish the policies needed to make a confident purchase

A shopper cannot evaluate an offer from product copy alone. Publish current shipping, return and refund, privacy, and terms information where it applies to your business. Make the rules easy to find and consistent with the checkout experience.

Useful

Specific return windows, conditions, shipping regions, processing expectations, costs, and contact routes.

Not enough

A policy link that returns an empty template, contradicts the product page, or omits a material purchase condition.

Eligibility rules are channel-specific. For example, Shopify's current ChatGPT storefront requirements list completed terms of service, privacy, and return and refund policies. Check the live Shopify requirements for the channel you plan to use.

Layer 4

Keep the open-web product evidence coherent

Browser agents and crawlers can still inspect the storefront separately from Shopify Catalog. Make representative product pages publicly reachable, linked in the sitemap, and consistent with the corresponding Shopify product record.

A practical Product JSON-LD review

On a representative product page, inspect the rendered Product structured data. Useful fields commonly include a name, image, description, brand, SKU or another product identifier, and an offer with price, currency, and availability. Validate the rendered page rather than assuming a theme setting produced correct markup.

Structured data helps machines interpret a page and supports eligible search features. It does not override Shopify Catalog, force a crawler to visit, or guarantee inclusion in a recommendation.

Robots and agent discovery

Check that /robots.txt does not unintentionally block the relevant public paths for the user agents you intend to allow. Robots rules are advisory instructions for requested open-web crawling. They are not authentication or a security boundary, and an allowed URL is not guaranteed to be crawled or indexed. Shopify says these rules do not disable data sharing through active Shopify Catalog channels.

Shopify automatically provides /agents.md as the canonical agent-discovery document. By default, /llms.txt and /llms-full.txt mirror it for compatibility. The managed document is enough for most stores, and none of these files replaces Shopify Catalog. Google also says llms.txt neither helps nor harms visibility or rankings in Google Search.

Only customize agents.md.liquid for an advanced requirement the managed file cannot cover. Doing so replaces Shopify's managed version and makes the merchant responsible for keeping discovery and commerce endpoints current. Avoid private contact information because the document is broadly cached and served to agents.

Use the Shopify Product JSON-LD checklist
Public-check boundary

Know what an external snapshot cannot see

A public storefront check can observe whether selected public paths responded, how sampled robots rules apply, whether agent-discovery documents are present, and whether one sampled product exposes basic Product fields.

It cannot read private Agentic admin settings, confirm channel activation or eligibility, inspect Catalog Mapping, audit Shopify Catalog ranking, or represent every product and variant. A path that could not be fetched should stay “not verified,” not be converted into a claim that the store is broken.

Working checklist

Fix the earliest broken layer first

  1. Access: review Agentic channel settings and current eligibility for the channel and market you care about.
  2. Retrieval: run realistic Catalog preview searches and inspect whether the right products appear.
  3. Meaning: improve weak titles, descriptions, categories, images, option names, identifiers, and live offer data.
  4. Mapping: map custom product sources when the useful facts are not in Shopify's default fields.
  5. Trust: complete and reconcile store policies, especially shipping and returns.
  6. Public evidence: sample several product types, check crawler access, sitemap discovery, and rendered Product JSON-LD.
  7. Measure: use Shopify's channel reporting and your own conversion data. Do not treat a technical pass as traffic or revenue.

Primary sources

Keep the checklist current

Shopify's agentic storefront capabilities are evolving. Recheck the official documentation before changing channel access or implementing a custom integration.

Catalog-wide follow-through

Move from one public sample to your Shopify product data.

AgentSight Free audits the 25 most recent products and labels the coverage it actually inspected.

Audit 25 products free