Schema That Actually Moves the Needle
Every ecommerce platform generates some form of Product schema markup by default. Shopify adds basic Product schema. WooCommerce has schema plugins. BigCommerce includes it natively. But the default implementations are minimal – they include just enough to trigger rich snippets in Google, not enough to power AI recommendations.
AI crawlers use Product schema differently than Google. They are not looking for snippets. They are building a structured understanding of your product to use in conversational recommendations. The properties that matter most are not the ones most implementations prioritize.
Tier 1: Must-Have Properties
These properties directly influence whether AI crawlers can recommend your product:
name
The product name. Sounds obvious, but many implementations use marketing titles instead of searchable product names. "The Ultimate Comfort Experience" tells an AI crawler nothing. "Nike Air Zoom Pegasus 41 Running Shoe" tells it everything. Use descriptive, specific product names in your schema.
description
This is the most underutilized property. Most implementations either leave it empty, stuff it with the same marketing copy from the page, or truncate it to one sentence. AI crawlers use the description to understand what the product does, who it is for, and how it compares to alternatives.
Write schema descriptions as if you are explaining the product to someone who cannot see the page. Include: what the product is, its primary use case, key differentiating features, and who it is best suited for. Aim for 100-200 words.
offers
The Offer object within Product schema must include:
- price – the actual current price
- priceCurrency – ISO currency code (USD, EUR, GBP)
- availability – use schema.org values: InStock, OutOfStock, PreOrder, etc.
- priceValidUntil – helps crawlers understand if pricing is current
- seller – your organization or store name
AI assistants frequently include price comparisons in recommendations. If your price is not in your schema, you are invisible in price-conscious recommendations.
aggregateRating
Review scores are one of the strongest signals AI crawlers use for product recommendations. Include:
- ratingValue – average rating (e.g., 4.3)
- bestRating – maximum possible rating (usually 5)
- ratingCount – number of ratings
- reviewCount – number of text reviews
Products with 4.0+ ratings and 50+ reviews are recommended significantly more often by AI assistants than products with no rating data.
Tier 2: High-Impact Properties
brand
Brand helps AI crawlers disambiguate products. "Pegasus 41" could refer to many things. "Nike Pegasus 41" is unambiguous. Always include a Brand object with the brand name.
review (Individual Reviews)
Beyond aggregate ratings, individual review objects give AI crawlers specific customer feedback to cite. Include 3-5 of your best reviews with author name, date, rating, and review body. AI assistants frequently quote from these when recommending products.
image
Include an array of product image URLs. While AI crawlers primarily process text, image URLs help with content verification and are used by multimodal AI models. Include at least one high-quality product image URL.
sku and gtin
Unique identifiers help AI crawlers match your product across databases. If your product has a GTIN (UPC, EAN, ISBN), include it. This is especially valuable for products sold by multiple retailers – the AI can verify your product is the same one reviewed on other sites.
Tier 3: Differentiating Properties
These properties set your schema apart from competitors running default implementations:
additionalProperty
Use PropertyValue objects to include specific product attributes: weight, dimensions, material, color, size range, battery life, compatibility. These properties power specific query matches. When someone asks "what is the lightest laptop under $1000," the AI needs weight data in structured form to answer.
isRelatedTo / isSimilarTo
Linking to related products within your schema helps crawlers understand your product catalog structure and can lead to multi-product recommendations.
award
If your product has won awards or certifications, include them. AI assistants mention awards when recommending products, as they serve as third-party credibility signals.
Common Product Schema Mistakes
- Duplicate schema across variants – every color/size variant has identical schema with only the SKU changed. AI crawlers see this as duplicate content.
- Schema on listing pages – putting Product schema on category listing pages where 20+ products are shown. This confuses crawlers about which product the schema describes.
- Outdated pricing – schema price does not match the current page price due to caching or dynamic pricing. This destroys trust signals.
- Fake reviews in schema – AI crawlers can cross-reference reviews across sources. Manufactured reviews in schema damage your credibility.
- Missing availability – if your product is out of stock and your schema says InStock, AI crawlers note the discrepancy.
Testing Your Product Schema
After implementing or improving your Product schema:
- Validate with Google's Rich Results Test for syntax errors
- Check that schema data matches visible page content exactly
- Monitor AI Visibility Scores for pages with updated schema
- Track AI crawler behavior on updated pages – do they crawl more frequently?
- Test AI assistants directly: ask ChatGPT or Claude about your product and see if the response uses your schema data
See what AI crawlers actually extract from your Product schema. Botjar shows you exactly which properties each crawler reads, which are missing, and how your schema completeness compares to competitors. Get your free bot audit →