When people shop through AI assistants, the answers depend on structured product data: what the product is, what it costs, whether it's in stock and which attributes it has. That data comes from Product and Offer schema on your pages and from merchant feeds you submit to platforms. Clean, complete, consistent product data is the technical foundation of AI shopping visibility.
Where AI shopping data comes from
- Product pages: visible content plus structured data
- Merchant feeds: Google Merchant Center and similar programs, including AI platforms' own merchant programs where available
- Retailers and marketplaces that list your products
- Reviews on your site and third-party platforms
The assistant matches a shopper's request ("lightweight waterproof hiking boot, wide fit, under $150") against those attributes. Missing or inconsistent data means you don't match. See ChatGPT shopping.
Product and Offer schema
A solid baseline for a single product:
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Trailhead 2 Waterproof Hiking Boot, Wide",
"description": "Waterproof leather hiking boot with a wide toe box and grippy outsole, for day hikes and light backpacking.",
"image": ["https://www.example.com/images/trailhead2-brown.jpg"],
"brand": { "@type": "Brand", "name": "Example Outdoor" },
"sku": "TH2-BRN-W",
"gtin13": "0123456789012",
"color": "Brown",
"material": "Full-grain leather",
"offers": {
"@type": "Offer",
"url": "https://www.example.com/boots/trailhead-2?variant=brown-wide",
"price": "149.00",
"priceCurrency": "USD",
"availability": "https://schema.org/InStock",
"itemCondition": "https://schema.org/NewCondition",
"shippingDetails": { "@type": "OfferShippingDetails", "shippingRate": { "@type": "MonetaryAmount", "value": "0", "currency": "USD" } },
"hasMerchantReturnPolicy": { "@type": "MerchantReturnPolicy", "merchantReturnDays": 30, "returnPolicyCategory": "https://schema.org/MerchantReturnFiniteReturnWindow" }
}
}
Include aggregateRating and review only for genuine reviews shown on the page.
Merchant feeds
Feeds deliver product data at scale and are required for many shopping surfaces. Key practices:
- Complete required attributes: ID, title, description, link, image, price, availability, brand, identifiers.
- Add optional attributes that matter for matching: color, size, material, pattern, age group, gender, product type, product highlights.
- Descriptive titles: product type plus key attribute ("Women's Waterproof Hiking Boot, Wide Fit, Brown").
- Update frequently: at least daily for price and availability, more often if they change quickly.
- Fix feed errors and warnings reported by the platform.
For AI platforms with their own merchant programs, follow their current specifications. OpenAI, for example, has published guidance for merchants who want their products to appear in ChatGPT shopping and support checkout.
Keep everything consistent
The same product should have the same price, availability, title and key attributes on:
- The visible product page
- Product schema
- Your merchant feeds
- Marketplace listings you control
Mismatches (a feed saying "in stock" while the page says "sold out") can lead to disapprovals and lost trust.
Variants
For products with sizes, colors or other options:
- Give each variant a unique identifier and, ideally, a URL or URL parameter that selects it.
- Provide variant-specific price, availability and images.
- Use a product group structure where the search engine supports it, following current guidance.
Attributes are your filters
AI shopping queries are full of constraints. Make sure attributes exist in text on the page, in schema and in the feed:
| Category | Attributes to cover |
|---|---|
| Apparel and footwear | Size, fit, color, material, care |
| Electronics | Compatibility, battery life, dimensions, ports |
| Home | Dimensions, materials, assembly, weight capacity |
| Beauty | Skin type, ingredients, fragrance-free, vegan |
| Food | Dietary attributes, allergens, servings |
See how to optimize product pages for AI assistants for the page-level copy.
Validate and monitor
- Run key product URLs through Google's Rich Results Test.
- Monitor Merchant Center diagnostics.
- Spot-check that schema renders in the server HTML. See JavaScript rendering and AI crawlers.
- Ask AI assistants shopping questions for your categories and see whether your products, with correct prices, appear.
Common product data problems
- Generic titles like "Trailhead 2" with no product type or attributes
- Missing identifiers (no GTIN, brand or MPN), making it hard to match the product across sources and reviews
- Prices in schema that lag behind the page after a sale ends
- Availability always "InStock" because the template hard-codes it
- Variants collapsed into one offer with a price range and no per-variant availability
- Ratings in markup that don't appear on the page, which violates guidelines
- Feeds updated weekly while prices change daily
Each one either stops a product matching a shopper's request or undermines trust in your data.
The bottom line
AI shopping runs on product data. Mark up products completely with Product and Offer schema, submit accurate, attribute-rich feeds, keep prices and stock consistent everywhere, and handle variants cleanly. Then the attributes shoppers ask for will match yours.
