AI Visibility Audit

The store was not cited in 5 of 5 tested buyer questions. Claude web search (Anthropic), run 4 November 2026, as a proxy for AI assistants. It is not ChatGPT, Google AI Mode or Perplexity; results vary by run and location.

If an assistant does not surface the store for a buyer's question, the buyer may never see it. The result is a snapshot from one tool.

Fix: Fix the data findings first, then repeat these questions monthly and track the trend rather than any single result.

Shopify Help: agentic storefronts and ChatGPT · StoreOps workflow

Sample only: 165 of 167 images (98.8%) have no alt text. Checked on 30 products spread evenly across the catalogue, because this field is not in products.json. This is not a full-catalogue count.

Alt text is the written description of an image. It helps people using screen readers and gives crawlers text they can read.

Fix: In the product's media, choose Add alt text. Describe what is in the image in a short plain phrase. Do not stuff keywords.

Examples: product-001, product-002, product-003, product-004, product-005, product-006, product-007, product-008, product-009, product-010

Shopify Help: image alt text · StoreOps workflow

6,298 of 6,357 products (99.1%) show no product type in the public feed. products.json does not expose Shopify's standard product category, so category could not be checked from public data.

Shopify says product types are not required, so this is not an error. Filled in, they still help with automated collections and filters, can feed Google's product_type attribute, and make the public catalogue easier for AI tools to read.

Fix: First check whether the standard Product category is already set in Shopify admin. Then add Type in bulk where it helps, starting with your biggest collections and one consistent list of types.

Examples: product-011, product-012, product-013, product-014, product-015, product-016, product-017, product-018, product-019, product-020

Shopify Help: product category and type · Shopify Help: product category · Google Merchant Center: product_type · StoreOps workflow

41 of 6,357 products (0.6%) have no images.

Google Merchant Center requires an image for each product, so these items cannot appear in shopping results or be shown properly by assistants.

Fix: Filter Products by those with no media, add a clear image of the actual item, or unpublish the product until you have one.

Examples: product-030, product-031, product-032, product-033, product-034, product-035, product-036, product-029, product-037, product-038

Google Merchant Center product data specification · Shopify Help: product details · StoreOps workflow

Sample only: 30 of 30 products (100.0%) have no barcode on any variant. Checked on 30 products spread evenly across the catalogue, because this field is not in products.json. This is not a full-catalogue count.

A GTIN helps Google identify a product made by another brand. Made-to-order or own-brand goods often have no GTIN, and that is allowed.

Fix: If you resell branded goods, copy the manufacturer's GTIN into the Barcode field. If you make the product yourself, leave it empty and tell Merchant Center it has no identifier. Never invent or buy codes to fill the gap.

Examples: product-011, product-001, product-002, product-003, product-004, product-005, product-006, product-007, product-008, product-009

Google Merchant Center: GTIN · Shopify Help: product details · StoreOps workflow

Return policy markup is missing on 3 of 3 sample product pages (the first 3 in products.json).

This states your return window in a form Google can read, so it does not have to interpret your policy page.

Fix: Add hasMerchantReturnPolicy to your Offer or Organization markup, or set returns in Merchant Center. The values must match your published policy.

Where: /products/product-011, /products/product-012, /products/product-065

Google Search: return policy structured data · Google Search: merchant listing structured data · StoreOps workflow

/llms.txt is present (760 words) but reads as platform-generated agent instructions rather than a description of this shop.

llms.txt is a proposal, not a ranking factor. This file gives agents instructions on how to shop, but says little about what makes this store different.

Fix: Optional and low priority. Check whether your platform lets you customise the file before spending any time on it.

Where: /llms.txt

llms.txt proposal · Shopify Help: agentic storefronts and ChatGPT · StoreOps workflow

107 of 6,357 products (1.7%) have an empty description or fewer than 50 words. 91 are empty and 16 are under 50 words.

AI tools and search engines can only repeat what the page says. A thin description gives them no size, material, use or delivery facts to quote.

Fix: Work through best sellers first. Add the real facts: what it is, size, material, how it is made and who it suits. Use only facts you can confirm.

Examples: product-039, product-040, product-041, product-042, product-043, product-044, product-045, product-046, product-047, product-048

Shopify Help: product details · Google Merchant Center product data specification · StoreOps workflow

220 of 6,357 products (3.5%) have no tags.

Tags power automated collections and on-site filters. They are an internal tool, so this matters less than the visible product data.

Fix: Agree a short tag list, then add tags in bulk from the Products page. Only do this if you will use them for collections or filters.

Examples: product-011, product-012, product-013, product-014, product-015, product-021, product-022, product-023, product-024, product-025

Shopify Help: product details · StoreOps workflow

483 of 86,775 variants (0.6%) have no SKU. 119 products are affected.

SKUs are your own stock codes. They do not affect buyers directly, but they make bulk edits, feeds and stock matching far safer.

Fix: Export products to CSV, fill the Variant SKU column with a simple consistent pattern and re-import a small test batch first.

Examples: product-011, product-012, product-013, product-014, product-015, product-026, product-027, product-028, product-016, product-029

Shopify Help: product details · StoreOps workflow

40 of 6,357 products (0.6%) share an exact title with another product. They fall into 18 groups.

When several products have the same name, buyers and machines cannot tell them apart, and they can compete with each other.

Fix: Review each group. Merge true duplicates into one product with variants, or add the detail that makes each title different.

Examples: product-051, product-052, product-053, product-054, product-055, product-056, product-057, product-058, product-059, product-060

Google Merchant Center product data specification · Shopify Help: product details · StoreOps workflow

25 of 6,357 products (0.4%) have a title or handle that looks like a test, copy or placeholder.

Test items and copies make a catalogue look unreliable and can surface in search. The match is on words such as 'test', 'copy' or 'sample', so some may be genuine products.

Fix: Review the list by hand. Set anything that is not for sale to Draft. Do not delete in bulk from this list alone.

Examples: product-030, product-031, product-032, product-033, product-034, product-035, product-036, product-016, product-049, product-050

Shopify Help: product details · StoreOps workflow

1 of 6,357 products (under 0.1%) have every variant unavailable.

Live product pages that cannot be bought are a dead end for a buyer who was sent there by an assistant.

Fix: Restock them, set them to continue selling if they are made to order, or unpublish them and redirect the URL.

Examples: product-064

Google Merchant Center product data specification · StoreOps workflow

3 of 6,357 products (under 0.1%) have at least one variant priced at 0.

A zero price is usually a mistake and can be rejected by shopping feeds. It may be intentional for a free add-on, so check each one.

Fix: Open each listed product, confirm the price and correct it, or hide the variant if it should not be for sale.

Examples: product-061, product-062, product-063

Google Merchant Center product data specification · StoreOps workflow

AggregateRating or Review markup is missing on 3 of 3 sample product pages. Review app code was found on the pages (judge.me). No review count was visible in the page HTML, so these products may simply have no reviews yet. Apps can also add markup with JavaScript, which this check does not run.

Review markup lets search engines read your rating. It is only valid when it reflects real reviews shown on that page.

Fix: If products have real reviews, turn on your review app's structured data setting, then test a reviewed product page. Never add ratings that are not on the page.

Where: /products/product-011, /products/product-012, /products/product-065

Google Search: review snippet structured data · StoreOps workflow

Shipping markup is missing on 3 of 3 sample product pages (the first 3 in products.json).

This states delivery cost and time in a machine-readable way for shopping results.

Fix: Add shippingDetails to the Offer markup, or manage shipping in Merchant Center. The values must match what checkout charges.

Where: /products/product-011, /products/product-012, /products/product-065

Google Search: merchant listing structured data · StoreOps workflow