Opens in a new tabSkip to content
Agent LighthouseAgent Lighthouse

    Searches the text of every published page. The evidence sources themselves are not in this index — search all of them on the trusted sources page.

    GitHub ↗
    Browse checks and page contents
    structured-data/service-schema

    Service schema

    What it checks

    AI agents use Service schema to understand what you offer and who provides it. Without it, agents must infer your offerings from unstructured text, which leads to inaccurate or incomplete descriptions in AI-generated recommendations.

    A Service or ProfessionalService node must carry name and provider. Product shapes are not this audit’s business any more — see the split below.

    Why it matters

    Google Merchant Center’s website-crawl, its automated-feeds path, crawls a server-rendered Product object that contains a nested Offer with price, priceCurrency, availability and condition, plus a sku or gtin. It uses that object to build and refresh the merchant’s catalog entries. Those entries populate the Shopping Graph that Gemini queries in Google AI Mode shopping and agentic checkout. It is not the ingestion path for ChatGPT shopping, Perplexity shopping, or Amazon Rufus, all of which are feed- or first-party-catalog-based.

    Evidence

    Product

    This is the single strongest schema.org signal in the AI era, and the strength comes entirely from Google’s commerce pipeline rather than from generic “AI reads schema” claims. Google documents the exact schema.org→attribute mapping (id→sku; gtin→gtin8/12/13/14/gtin/isbn; price→price+priceCurrency; availability→availability; condition→itemCondition) and states: “If you’re using automatic item updates, make sure to specify the schema.org properties price, priceCurrency, availability, and condition” (google-merchant-supported-structured-data). Setup requires a Product object with a nested Offer (google-merchant-setup-structured-data).

    Automated feeds are built by “website crawl”, which “uses structured data and sitemap information to extract the most up-to-date information about relevant products”, re-checked at least every 24 hours (google-merchant-automated-feeds). Offer-level SKU/GTIN annotation is what lets the crawler match page offers to catalog items. Google’s Product doc confirms the hybrid model. “Providing both structured data on web pages and a Merchant Center feed maximizes your eligibility to experiences.” And “product snippets may use pricing data from your merchant feed if it’s not present in the structured data on the page” (google-product-structured-data).

    Downstream, AI Mode shopping “brings together Gemini capabilities with the Shopping Graph”, which holds 50B+ listings with reviews, prices, colors and availability, 2B refreshed hourly (google-shopping-graph-ai-mode). Google’s changelog also shows Product markup under ACTIVE development in 2026 (Product.category codes, a new “Sale duration” section, hasAdultConsideration) while FAQ and HowTo were being retired (google-search-updates-changelog). Hard audit rule from Google: “Structured data markup must be present in the HTML returned from the web server. The structured data markup can’t be generated with JavaScript after the page has loaded.”

    Limits

    OpenAI’s product feed spec is feed-only. Required fields are item_id, title, description, brand, url, image_url, availability, price, is_eligible_search, is_eligible_checkout, seller_name and target_countries. GTIN is optional. There is no mention of schema.org, JSON-LD or on-page markup anywhere in the spec or the wider commerce docs index (openai-product-feed-spec, openai-commerce-docs-index). The Agentic Commerce Protocol feed spec likewise defines its own taxonomy and does not reference schema.org, GTIN, or Merchant Center compatibility (acp-feed-spec). Perplexity’s merchant program reportedly requires CSV/XML feeds (perplexity-merchant-program-terms — UNVERIFIED, hub returns 403).

    Amazon Rufus draws on the Amazon catalogue, customer reviews, community Q&A and “public information on the web”, with no mention of schema.org or merchant markup crawling (amazon-science-rufus). And Google’s AI guidance still says “there’s no special schema.org markup you need to add”, routing merchants to Merchant Center feeds and Business Profiles instead (google-ai-optimization-guide). Honest framing for the dossier. Product markup earns its A because it is a documented ingestion path into a Google system that demonstrably powers an AI shopping surface. It does not earn it because an LLM reads it off the page at answer time.

    How it scores

    The strongest schema.org signal in the corpus, and its strength comes from one specific pipeline rather than from a general “AI reads schema” claim. Google documents the exact schema.org-to-attribute mapping used by Merchant Center’s website crawl: id to sku, gtin to the GTIN family, price to price and priceCurrency, and availability and condition each to their own attribute. Those entries feed the Shopping Graph that Gemini queries in AI Mode shopping. A named consumer with a published field mapping is grade A. It does not generalise to OpenAI, whose product feed specification is feed-only and never mentions schema.org, JSON-LD or on-page markup.

    Sources