Shopify AI visibility checker: how to test one product

What does a Shopify AI visibility check actually tell you?

A Shopify AI visibility check can mean two different things: inspecting the product data a tool can read, or observing whether an AI shopping answer includes the product. A page scan answers the first question. Only a captured shopping test can answer the second for its recorded question, market, system and date.

MC Lab Self-Check combines a free Product Data Preview with an optional paid, recorded AI shopping diagnostic for one public Shopify product URL. The preview checks observable product data and access. The paid diagnostic records the shopping conversation, sources and observed point where the product enters or leaves the decision.

Check your Shopify product page for free

The free preview does not run an AI recommendation test or use credits. The paid diagnostic is a controlled API observation; it is not a measurement of your ranking in the consumer ChatGPT interface.

Choose the check that answers your question

“Can this tool read my product page?” — Free Product Data Preview

Submit one public Shopify product URL. Review the title, category, price, currency, availability, structured data and access information the preview returns. Check that these describe the intended product and market. Missing or conflicting fields identify a source question to investigate; they do not establish why a shopping AI omitted you.

“Does my product enter a shopping conversation?” — Paid Self-Check

After the preview, select the target market and run the recorded diagnostic using credits. Review the actual questions, answers, cited sources and target outcome. MC Lab selects the model configuration internally. Read the conditions in the report before interpreting its result or comparing it with another run.

“What should we change, and why?” — Professional review diagnosis

The professional Product Candidate-Set Diagnostic uses a more carefully designed AI workflow, more compute and more capable AI to investigate the failure, compare competitor evidence and develop a focused change-and-retest plan. It is useful when the observed omission or inaccurate comparison warrants deeper investigation. It is a separate service from software credits.

How to check your product in ChatGPT or another shopping assistant

You can make a small manual observation before buying a tool. Choose a search-enabled consumer experience, record its visible mode, and keep this evidence separate from any automated API report.

1. 1

   Choose one product, buyer need and market. Write down the decision a plausible shopper is trying to make. Include only constraints the buyer would naturally know.

2. 2

   Start a fresh conversation without naming your product. Do not paste the product URL, brand, distinctive marketing copy or a list of unique specifications into the shopping question.

3. 3

   Save the complete answer and sources. Record the exact question, date, mode, market and follow-ups. Capture competing products and reasons given, including outcomes you did not expect.

4. 4

   Check direct retrieval in a separate conversation. Give the product URL and ask the system to identify the product from that page. Verify the correct SKU and source grounding. Keep this named control out of the unbranded shopping conversation.

5. 5

   Classify the observation before proposing a fix. Was the target omitted, included, compared inaccurately, rejected on a valid trade-off, or recommended? A failed or blocked run is incomplete evidence.

A natural buyer question

I'm shopping in the US for a compact everyday backpack for commuting by train. I carry a laptop and want something comfortable that does not feel oversized. What should I consider?

This is an illustrative prompt, not a captured result. Use a buyer need that fits your own product. Avoid adding highly distinctive requirements merely to make your product the only possible answer. If you ask follow-ups, let them resolve ordinary purchase questions.

What does a missing AI mention mean?

The page is readable, but the product is absent

The scan and the shopping test answered different questions. Direct access shows that information can be read under the control conditions. It does not show that the system retrieved or considered that product during the unbranded conversation. Compare the buyer's language, category associations and cited competitor evidence before deciding on a change.

The product appears, but another option is preferred

Check whether the comparison is accurate. A competing product may better fit the buyer's budget, size requirement or delivery need. Candidate-set inclusion and final ordering are separate outcomes; being listed second is not automatically a problem.

The system cannot correctly identify the product by URL

Resolve the retrieval uncertainty first. The page could be unavailable to that system, confused with another variant, or insufficiently grounded in the returned sources. A timeout or tool failure must not be recorded as proof that your product lacks demand or visibility.

The product is included, but its facts are wrong

Save the incorrect statement and cited source. Compare the current product page, variants, shipping conditions and relevant outside references. Correct verified inaccuracies where you have control, then repeat the corresponding check after the changed information can be read externally.

Does Shopify Catalog inclusion mean AI will recommend my product?

No. Shopify states that Catalog inclusion does not guarantee a particular AI answer or position. Catalog data gives supported channels information they can use; the channels control their own results. See Shopify's Catalog documentation.

Shopify recommends accurate descriptions, visible product essentials and comprehensive product information in its guide to optimizing a store for AI. Treat this as maintaining trustworthy product evidence, then test how that evidence is used for a particular buyer need.

Will adding schema or llms.txt guarantee AI visibility?

No. Google says that its AI Overviews and AI Mode use the existing SEO foundations and require no special AI files or schema. A supporting page still needs to be indexed and eligible for a search snippet. See Google's AI search guidance. Those requirements concern Google Search; they are not a universal specification for every assistant.

Does a ChatGPT crawler setting also control training?

OpenAI distinguishes OAI-SearchBot for search from GPTBot for training. Their controls serve different purposes. Consult OpenAI's crawler documentation before changing access. Permitting a crawler is an access decision, not evidence that it has cited or recommended you.

What should I do after the check?

Keep a record of the buyer question, tested surface, target market, product outcome, source URLs and unresolved uncertainty. If the comparison was accurate and useful, there may be nothing to change. If an addressable source problem exists, choose the smallest useful correction and retain the original record for comparison.

A comparable retest uses the same buyer need, market and relevant conditions. Confirm that the changed information is externally readable first. Record what changed in the answer without treating one result as a permanent ranking or a sales forecast.

For a lightweight first step, use the free preview. If you need recorded shopping evidence, Self-Check offers one-time credit packs from USD 9; see current packs and credits. There is no subscription required. Professional review diagnosis is available separately when you need a deeper investigation.