AI Visibility for Online Stores: Products, Feeds, Schema, Reviews, and Availability
An AI system can only recommend a product if it can determine what you sell, at what price, in which variant, with what availability, and who it's suitable for. The top priority for an online store therefore isn't generic GEO articles, but accurate product data, crawlability, feeds, structured data, and trust.
What shopping queries look like
Users add conditions such as:
- running shoes under $80 for a wider foot;
- a black work jacket available in XL;
- a coffee machine for a small kitchen with automatic milk frothing;
- a local store with fast delivery;
- an alternative to a specific product.
The system needs to filter by attributes. If those parameters exist only in an image or marketing copy, comparison becomes much harder.
The product page
It should include:
- an unambiguous name;
- brand and model;
- GTIN, MPN, or SKU as appropriate for the category;
- variant, color, and size;
- material and dimensions;
- compatibility;
- price and currency;
- availability;
- delivery information;
- returns and warranty claims;
- an original description;
- photos of the actual product;
- customer questions;
- reviews, provided they're authentic.
Supplement the manufacturer's description with your own experience: who the product suits, how the sizing runs, what the color looks like in reality, or what it's compatible with.
Google Merchant Center
Google recommends using Merchant Center and quality feeds for products. Keep these consistent in particular:
- a stable ID;
- the product and image URLs;
- a name free of spam;
- an accurate description;
- price;
- availability;
- brand and identifiers;
- shipping;
- return policy;
- variants.
The feed and the page must not contradict each other. A mismatched price or availability status undermines the reliability of your data.
Product and Offer schema
JSON-LD can explicitly describe a product and its offer:
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Product model",
"sku": "SKU-123",
"brand": {"@type": "Brand", "name": "Brand name"},
"offers": {
"@type": "Offer",
"url": "https://example.com/product",
"priceCurrency": "USD",
"price": "129.00",
"availability": "https://schema.org/InStock"
}
}
Generate it from the same data source as the visible page. Schema doesn't guarantee an AI recommendation, but it can help Search understand the offer and unlock supported product-rich results.
Variants and canonicalization
Clarify:
- whether each variant has its own indexable URL;
- how price, image, and availability change between variants;
- which canonical is the correct one;
- whether a variant is accessible without user interaction;
- whether internal links to variants are crawlable.
Endless filter combinations can create a crawl trap. Only index pages that carry real search and user value.
JavaScript and accessibility
Core product data must not be locked behind a click or a login. Check both the raw HTML and the rendered version for:
- name and price;
- availability;
- description;
- canonical;
- structured data;
- links to variants.
Reviews and reputation
Your own reviews help customers, but independent sources can confirm the store's overall trustworthiness. Relevant options include comparison sites, your Google Business Profile, and industry publications.
Don't buy ratings or reuse someone else's reviews without permission. Review markup in schema must comply with the target search engine's rules.
ChatGPT and product discovery
OpenAI supports product discovery in ChatGPT and publishes information for merchants. Follow the current documentation on feeds and supported capabilities, since product features change quickly. Don't assume that ordinary crawling can substitute for an accurate commerce feed.
Prioritizing a large catalog
Start with:
- bestsellers;
- the highest-margin products;
- high-demand categories;
- products with unique data;
- pages with price or availability errors.
For everything else, at minimum ensure a consistent feed, schema, and basic description.
Checklist
- Products are accessible to crawlers.
- The feed matches the page.
- Price and availability are current.
- Variants have correct URLs and canonicals.
- Product/Offer schema is valid.
- Identifiers are filled in.
- Descriptions add original value.
- Images match the variant.
- Shipping and returns are clearly stated.
- Reviews are authentic.