Schema for AI search: the fields that matter.

Why schema still matters in an AI-answer world

An AI engine that cites a product needs to be confident about what the product is. Structured data is how a PDP tells that story in machine-readable form. In the blue-link world, schema drove rich results — stars, prices, availability ribbons under the link. In the AI-answer world, schema drives citation confidence: an engine that can resolve "Ravenback Founder Parka" to a specific SKU with a specific price, GTIN, and material is more likely to cite it than one that can only see a marketing headline.

The Schema.org Product vocabulary has not changed much in years. What has changed is which fields AI engines actually use. eCommerce Insights's audits of Shopify PDPs in Q1 2026 (illustrative, across thousands of stores) show that complete Product blocks correlate with higher citation share across ChatGPT, Perplexity, and Google AI Overviews. Correlation is not cause, but the direction is consistent.

The Product schema fields AI engines use most (Q1 2026)

Not every field on the Schema.org Product type earns its keep. The following fields carry the most weight based on observed citation behavior and eCommerce Insights's manual engine-response reviews. The list is ordered roughly by impact.

name and description

The name field is the product's canonical title. Include the brand, the product model, and the key variant attribute. "Ravenback Founder Parka — Men's, Slate" beats "Founder Parka." The description is a short factual summary, usually two to four sentences, that an engine might quote directly. Keep it citable: material, use case, dimensions or sizing, and one differentiator. Avoid marketing adjectives.

sku, gtin13, mpn, brand

These four identifiers let engines resolve the product as a distinct entity. sku is the merchant's internal identifier — Shopify's variant SKU fills this. gtin13 is the global trade item number (UPC in the US, EAN in Europe). mpn is the manufacturer part number, useful for brands that manufacture their own goods. brand is a nested Brand object with the brand name. Filling all four is the strongest single signal of catalog legitimacy an AI engine reads.

offers (price, priceCurrency, availability, priceValidUntil)

The offers object is how pricing data reaches engines. Required fields: @type "Offer", price as a string, priceCurrency as a three-letter ISO code, and availability as a Schema.org URL like https://schema.org/InStock or OutOfStock. Useful additions: priceValidUntil, which signals engines that the price is fresh, and itemCondition for used or refurbished SKUs. eCommerce Insights audits in Q1 2026 find priceValidUntil missing on roughly 80% of Shopify PDPs (observed sample, illustrative).

Fill the identifiers once. An engine that can resolve your SKU is an engine that can cite it with confidence.

image

The image field is an array of image URLs. Engines use these for grounding and for the visual tile that shows inside AI shopping surfaces. Include at least three images: the primary product shot on a plain background, a lifestyle or in-use shot, and a detail shot. Each URL should resolve to a full-size image, not a thumbnail. Shopify's product images can be referenced at their canonical CDN URLs.

additionalProperty (material, pattern, color, size, country_of_origin)

This is the most underused field in Shopify schema. additionalProperty takes an array of PropertyValue objects, each with a name and value. Use it to describe material composition, color, pattern, size, country of origin, warranty length, weight, and any other spec that does not map cleanly to a top-level field. An AI engine answering "what is this jacket made of" can cite a PDP that exposes material as structured data far more readily than one that buries it in a paragraph.

aggregateRating and review

If the PDP has real reviews — twenty or more, ideally — the aggregateRating block is a strong citation signal. It requires ratingValue, reviewCount, and usually bestRating. Adding review entries (individual review objects with author, rating, and reviewBody) deepens the signal further. Do not fabricate this data — a review count out of sync with the visible review widget is a credibility drop and, as of Q1 2026, triggers engine skepticism on both Google and ChatGPT.

category

The category field takes a string describing the product's taxonomy position. Google Product Taxonomy strings work well: "Apparel & Accessories > Clothing > Outerwear > Coats & Jackets." Shopify's native theme often leaves this blank because Shopify categories and Google categories do not map automatically. Filling it helps engines place the product in a shopping context.

isRelatedTo / isSimilarTo

These fields link to related products as URL references. They are low-cost to add for themes with a "you might also like" block. Engines use them to build a product graph — knowing that the Founder Parka is similar to the Terminal Trench helps with comparison-intent citations. Use sparingly; three to five related products per PDP is plenty.

The common blanks on Shopify PDPs

Across the thousands of Shopify PDP audits eCommerce Insights has run, the pattern of missing fields is consistent. The most common blanks, in order of frequency: gtin13 (missing on 70%+ of PDPs, illustrative), mpn, priceValidUntil, additionalProperty entirely, category, isRelatedTo, and weight on physical goods. Filling these is a day of theme work or a few hours with a schema app and compounds across every SKU in the catalog.

The full Product JSON-LD example

Below is an illustrative, complete Product JSON-LD block for a fictional apparel SKU. Copy, adapt, replace the values with real product data.

<script type="application/ld+json">
{ "@context": "https://schema.org", "@type": "Product", "name": "Ravenback Founder Parka — Men's, Slate", "description": "A 650-fill down parka in a water-resistant recycled-nylon shell, cut for layering. Adjustable hood, four-way storm cuffs, insulated hand pockets. Designed in Portland, Oregon. Unisex sizing, runs true to size.", "sku": "RV-FP-M-SLATE", "gtin13": "0810123456789", "mpn": "RV-FP-2026-M-SLATE", "brand": { "@type": "Brand", "name": "Ravenback" }, "category": "Apparel & Accessories > Clothing > Outerwear > Coats & Jackets", "image": [ "https://ravenback.com/cdn/shop/products/founder-parka-slate-front.jpg", "https://ravenback.com/cdn/shop/products/founder-parka-slate-back.jpg", "https://ravenback.com/cdn/shop/products/founder-parka-slate-lifestyle.jpg" ], "offers": { "@type": "Offer", "url": "https://ravenback.com/products/founder-parka?variant=m-slate", "priceCurrency": "USD", "price": "485.00", "priceValidUntil": "2026-12-31", "availability": "https://schema.org/InStock", "itemCondition": "https://schema.org/NewCondition", "seller": { "@type": "Organization", "name": "Ravenback" } }, "additionalProperty": [ { "@type": "PropertyValue", "name": "material", "value": "Recycled nylon shell, 650-fill responsibly-sourced down" }, { "@type": "PropertyValue", "name": "color", "value": "Slate" }, { "@type": "PropertyValue", "name": "size", "value": "Men's M" }, { "@type": "PropertyValue", "name": "country_of_origin", "value": "Vietnam" }, { "@type": "PropertyValue", "name": "weight", "value": "1.2 kg" }, { "@type": "PropertyValue", "name": "care", "value": "Machine wash cold, tumble dry low" } ], "aggregateRating": { "@type": "AggregateRating", "ratingValue": "4.7", "reviewCount": "124", "bestRating": "5" }, "isSimilarTo": [ "https://ravenback.com/products/cascade-field-jacket", "https://ravenback.com/products/terminal-trench" ]
}
</script>

Where to put the schema on a Shopify theme

Shopify themes emit Product JSON-LD from a partial named product-json-ld.liquid or equivalent, typically included in sections/main-product.liquid. Dawn, Sense, Refresh, Crave, and most Horizon-generation themes all include a base schema block. Apps like Schema App, JSON-LD for SEO, or Yoast SEO for Shopify overlay an additional block.

Best practice: audit the live PDP, decide which block is more complete, keep that one, and disable the other to avoid duplicates. eCommerce Insights's product schema generator flags duplicate blocks automatically. The PDP optimization workflow walks through the diff-and-push cycle.

Validation: Google's Rich Results Test and beyond

Start with Google's Rich Results Test for syntactic correctness. It catches missing required fields and malformed JSON. Pass Schema.org's official validator for coverage completeness. Use the AEO grader to measure how the schema performs against AI-specific citation signals — the Rich Results Test says the schema is valid; the AEO grader says whether it is competitive.

Run validation after every theme update, after every app install, and after any manual change to the schema Liquid partial. Small breaking changes compound across a catalog fast.

A schema block that validates is a floor. A schema block that gets cited is a ceiling only complete data earns.

Key takeaways

• Complete Product JSON-LD is the highest-leverage technical move.
• The fields that matter most: name, description, sku, gtin13, offers, image, additionalProperty, aggregateRating (if real).
• The common Shopify blanks are gtin13, mpn, priceValidUntil, and additionalProperty.
• One clean block beats two conflicting ones — deduplicate before adding more.
• Validate after every theme or app change; small breakage compounds.

Related guides

Tools and product

• Product schema generator — audit and generate complete Product JSON-LD for any Shopify PDP.
• The eCommerce Insights product — schema recommendations as diffs, human-approved, one-click push to Shopify.