September 26, 2026 · The lunalink.ai team

Should Shopify merchants check policy pages before asking AI shopping tools to recommend products?

Hands typing on a laptop showing an e-commerce website in a modern office setting. — Photo by Shoper .pl on Pexels

A customer may now find your product without first reading your homepage, collection page, or FAQ.

They might ask an AI shopping tool for a product that meets a need, price range, or delivery deadline. The tool may use structured catalogue data, public pages, and store information Shopify provides to AI channels.

That makes policy pages part of the buying path. Not the most exciting part, but a real one.

Before you spend time asking whether an AI tool can recommend your products, check whether your store explains what happens after someone buys.

Why policy pages matter for AI shopping

Shopify’s agentic storefront tools include Shop policy completeness as a ranking signal. That covers shipping, returns, refunds, and other store policies. Shopify says policies can help show that a store is legitimate and help AI channels recommend products confidently.

This is sensible from the customer’s side.

Say your product is the right size, colour, and price. An AI tool may still need to understand whether it can be delivered where the customer lives, what delivery costs, and whether it can be returned. A product recommendation that skips those details can lead to a poor purchase.

Shopify also sends structured product information through Shopify Catalog. That can include titles, descriptions, options, images, prices, and availability. An AI channel does not always need to browse your store as a person would.

So do not assume a beautiful storefront fills gaps in unclear policies.

Your public store information needs to agree with your catalogue data and with the experience a buyer gets at checkout.

This is not only an SEO tidy-up

For stores on Shopify’s Agentic plan, published policies are part of the setup requirements before selling through agentic storefronts. Shopify requires:

A terms-of-service policy
A privacy policy
A return-and-refund policy

That is more direct than general advice to keep your site trustworthy. It is a setup requirement for that plan.

The exact requirement depends on your plan and channel. Shopify’s wider agentic storefront documentation says eligible stores can have products discovered in ChatGPT, Google AI Mode, Gemini, Microsoft Copilot, and Meta. But the separate Agentic-plan setup page is the page that explicitly lists the policy requirement before selling.

Check the rules that apply to your own Shopify plan. Do not borrow a checklist from another merchant and assume it matches your setup.

Start with a one-minute policy check

In Shopify admin, open Settings → Policies.

Look at each published policy as if you were a customer who has never bought from you. You are checking for missing facts, not trying to make the page sound impressive.

For shipping, check:

Where you ship
Whether there are places you do not ship
How shipping cost is calculated
When an order is expected to leave your warehouse
Whether delivery timing is an estimate or a guarantee

For returns and refunds, check:

How long a customer has to request a return
Which items cannot be returned
Whether return shipping is paid by you or the customer
When refunds are issued
Where the refund goes

For privacy and terms, check that they are published, linked, and still describe how your store works.

If a sentence leaves you asking “what happens next?”, a buyer or AI tool may have the same problem.

A policy can be short. It just needs to be true and specific.

“Returns accepted” is incomplete. “Returns accepted within 30 days of delivery for unused items in original packaging; customers pay return postage” gives someone a usable answer.

Only publish terms you can actually follow. Clear but inaccurate policy text creates a different problem.

Check what Shopify makes available

Shopify automatically serves discovery files for every store:

`/agents.md`
`/llms.txt`
`/llms-full.txt`

By default, these files include store policies alongside the store name, store URL, sitemap, and other discovery endpoints.

You can check this yourself in under a minute. Add one of those paths to your store domain in a browser.

For example:

`https://yourstore.com/agents.md`

Then search the page for the word `policy`.

We built our own free scanner to read public URLs only: robots.txt, the catalogue endpoint, product markup, and policy pages. No install or login is needed to learn the basics. That decision came from a simple constraint: if public information is enough to spot a gap, a merchant should not need to trust a stranger’s app first.

The same principle applies here. Open the public files yourself. Check that the policy links lead to live pages. Check that those pages say what you think they say.

Do not treat the discovery file as a replacement for the policy page. It is a way to see what an AI agent may be pointed toward.

Keep product promises and policy terms together

A common problem is not a missing policy. It is a mismatch.

A product page might say “fast delivery,” while the shipping policy gives a longer dispatch window. A product description might say “easy returns,” while the refund policy excludes that product category. A sale page might imply a refund, while the policy only offers store credit.

Those are customer-service problems first. They also make the information around your product harder to trust.

OpenAI says its shopping ranking can consider availability, price, quality, whether a merchant is the primary seller, and whether Instant Checkout is enabled when several merchants sell the same product. Product relevance is not the only consideration.

OpenAI’s commerce policies also apply to linked merchant webpages, as well as product feeds and listings. They prohibit deceptive or misleading activity, including unclear or unfair terms, hidden fees, and misleading return or refund policies.

You do not need to write policies for an AI system. Write them for the person receiving the order.

That usually produces the same thing AI shopping channels need: accurate, consistent information.

Do not rely on robots.txt as a policy strategy

You may have chosen to block some AI crawlers in `robots.txt`. That is a separate decision from keeping public policies accurate.

Shopify says robots.txt rules are advisory and affect open-web discovery only. They do not stop Shopify Catalog product data from reaching activated agentic storefronts.

If Shopify Catalog access is turned off, products may still be found through ordinary web crawling and indexing. Shopify warns that information found that way may be less complete, accurate, or current.

So there is no useful version of this where unclear public policies stop mattering.

If your store is eligible for agentic storefronts, also check Settings → Apps and sales channels → Agentic storefronts in Shopify admin. Shopify says stores are opted in by default to “Allow Shopify to manage for me.” That setting can provide catalogue access to AI channels, enable direct checkout where supported, and add new agentic storefronts automatically.

You can switch to individual channel controls if that better fits your store. But first understand what is currently enabled.

A practical order of work

If you are getting your store ready for AI shopping discovery, do this in order:

1. Open Settings → Policies and make sure your required pages are published. 2. Read shipping, return, and refund terms against your real operations. 3. Compare policy wording with product pages, collection pages, and checkout messages. 4. Visit `/agents.md` on your live domain and confirm policy links work. 5. Check your Agentic storefronts setting and the requirements for your Shopify plan.

This is not a promise that a policy edit will produce recommendations. Shopify and AI channels use more than one signal, and product data still needs to be accurate.

But it removes a basic gap. A shopper should be able to learn what they are buying, what it costs, when it arrives, and what happens if it is wrong.

Start by reading your return policy on a phone. If you cannot understand it without opening three more pages, fix that before asking an AI shopping tool to send customers there.

Sources

shopifyai shoppingstore policiesproduct data