AI visibility for
ecommerce.
Give shoppers stable product facts, clear delivery conditions and accessible category pages rather than promises of guaranteed AI recommendations.
Product facts are the starting point
A shopping question often includes material, size, compatibility, budget or delivery location. Make those facts visible on the product page and consistent with the catalog. A polished image does not substitute for dimensions or a readable compatibility table. Explain variants without sending the reader through a JavaScript-only selector for every essential attribute.
Prioritize useful pages
- Publish actual specifications, unit conventions and what is included in the box.
- State delivery regions, dispatch expectations and return conditions with a revision date.
- Keep category descriptions useful and avoid indexing every filter combination.
- Reconcile product availability and prices across visible content and structured data.
Schema without invented signals
Use Product for a real item and Offer for its current sale conditions. An illustrative description might say: “This [material] desk is [verified width] wide and ships flat-packed to [supported region].” Replace placeholders before publication. Mark up reviews only when they are genuine and actually displayed; schema does not make a fabricated rating acceptable. Canonical URLs need to match the intended variant and indexing strategy.
Inspect citations as well as mentions
Sample a broad category, a constrained buying need and a brand-policy question separately. A citation to an old price or reseller listing may explain an inaccurate answer. Keep the source URL and answer date together. The local mock fixture cannot establish stock, suitability or shopping visibility for a real store.
Technical failures to check on online stores
- Variant prices loaded after render. Size and colour selectors that fetch prices with JavaScript leave the initial HTML without a price. Render the default variant's price in HTML and keep
Offermarkup in step with it. - Faceted URLs competing with category pages. Filter combinations create thousands of thin URLs. Canonicalise them to the main category, and keep the sitemap to pages you want found.
- Out-of-stock products still marked in stock. Stale
availabilityin JSON-LD contradicts the visible page. Generate markup from the same data as the product template so both change together. - Shipping and returns only in a footer link. Delivery conditions often sit on one policy page. Summarise delivery time, cost and returns window on product pages, and link the full policy.
- Platform robots.txt left on defaults. Shopify and other platforms ship a default robots.txt that may not match your AI crawler policy. Review it and make an explicit choice for search and training crawlers.
Pick one product with several variants. Fetch the URL with curl and confirm the default price and availability are in the HTML, then compare them with the Offer in the JSON-LD. Mark a test product out of stock in your admin and check that both the visible badge and the markup change. Count the URLs in your sitemap against the number of live products and categories; a sitemap that is many times larger usually contains filter or parameter URLs. Finally, read the robots.txt your platform serves rather than assuming the defaults suit you.
Example JSON-LD
Use Product with a nested Offer on each product page. Every value should match what the shopper sees on the same page. The example below is illustrative: replace every bracketed value with facts from your own page.
{
"@context": "https://schema.org",
"@type": "Product",
"name": "[Product name]",
"sku": "[SKU]",
"brand": {
"@type": "Brand",
"name": "[Brand]"
},
"image": "https://[domain]/images/[product].jpg",
"description": "[Short factual description]",
"offers": {
"@type": "Offer",
"url": "https://[domain]/products/[slug]",
"price": "[Price]",
"priceCurrency": "[Currency]",
"availability": "https://schema.org/InStock"
}
}Add aggregateRating only when the reviews are visible on the page and collected according to the review platform's rules.