AI search fix
Schema markup for AI search: Organization, Product, FAQPage and Breadcrumb
Schema markup — schema.org JSON-LD — helps AI systems and search engines read facts on your pages without guessing from messy HTML: who you are (Organization), what you sell (Product and Offer), your real FAQ content (FAQPage), and where a page sits in your site (BreadcrumbList). Every field must match what a visitor actually sees; markup that states a price, rating or question no human can find on the page is misleading and can violate spam policies. Google does not require structured data for generative AI features in Search, and no schema type guarantees a citation in ChatGPT, Perplexity or AI Overviews — it only reduces the chance a parser misreads or misquotes your page.
Fix crawl access before investing in schema: if GPTBot or another AI crawler receives a 403, parsers never see your JSON-LD in the first place. Add markup after bots can already fetch your pages, not instead of fixing access.
Core schema types for AI-readable pages
| Type | What it does and where to use it |
|---|---|
| Organization | Names your business, canonical URL and logo — homepage or a canonical about page, one block per site. |
| Product / Offer | Price, availability and brand for a single item — one focused product or plan page per block. |
| FAQPage | Question/acceptedAnswer pairs — only on pages with a real, visible FAQ section. |
| BreadcrumbList | Ordered hierarchy of a page's position — must match your real, crawlable navigation. |
Free · 2 minutes · no card
See what AI crawlers hit on your site
Technical blockers, missing context, weak AI-readiness signals — in one HTML report.
You'll get an HTML report on Organization, Product, FAQPage and BreadcrumbList coverage on your live pages.
Organization schema: one consistent brand entity
Organization JSON-LD names your business, canonical website, logo and any sameAs profiles you actually link to — publish it on the homepage, or a canonical about page if that is where your brand lives, and avoid conflicting Organization blocks on every template. Match every field to what your footer or contact page already shows: legal or trading name, an https URL, a logo URL and a contact method, then set @id to your homepage URL plus #organization so other schema on the site can reference the same entity.
Pair Organization markup with crawl readiness first: if GPTBot receives a 403 at your CDN, parsers never reach the JSON-LD regardless of how well it is written. Validate in view-source and a rich-results checker after every deploy, since theme or plugin updates can silently duplicate or break the block.
Product schema: match Offer to what shoppers actually see
Product schema describes what you sell — name, image, brand and usually a nested Offer with price, priceCurrency and availability. Use it on URLs that show a single primary item, not bundle or category pages listing dozens of SKUs, and copy the visible title, price and image exactly as shown to shoppers; never mark up items hidden behind login or out of stock unless the page itself says so.
Validate the live HTML after every CMS publish — Shopify and WooCommerce plugins are prone to serving stale, cached JSON-LD that no longer matches the current price. Only mark up aggregateRating if the reviews are genuinely visible and compliant; fabricated ratings mislead both users and AI parsers and can violate spam policies.
FAQPage schema: only for questions a visitor can actually read
Google stopped showing FAQ rich results in Search from May 2026, so FAQPage markup no longer buys a SERP feature — but a visible, accurate FAQ section with matching JSON-LD can still help AI systems parse your questions and answers cleanly. The rule that matters most has not changed: mark up only questions and answers that exist, word for word, in the page's visible HTML. A Question with no matching on-page text is misleading structured data, not a shortcut.
Boilerplate FAQ blocks copied across an entire site are a common failure mode — they conflict with each page's real intent and often duplicate a second FAQPage block emitted by a theme or plugin. Keep one focused FAQ section per relevant page, remove duplicates, and revalidate with a rich-results tool after any CMS or plugin update.
BreadcrumbList: match real navigation, then validate it
BreadcrumbList JSON-LD needs ordered ListItem entries with position starting at 1 and absolute, canonical item URLs — and it has to match your site's real, crawlable navigation. The most common mistake is a trail that adds a category level no visible link points to, or a duplicate BreadcrumbList block where a theme template and a plugin each emit a conflicting hierarchy on the same page.
Before publishing, confirm every item URL resolves with an HTTP 200, not a redirect or a 404 — a common regression after a URL migration that updated visible links but left the old JSON-LD in place. Re-check breadcrumbs whenever navigation changes, and spot-check a handful of URLs per template type, since a category template and a product template often build breadcrumbs from different logic.
Frequently asked questions
Product vs Offer — which do I need?
Product describes the item itself; Offer describes the commercial terms. Most e-commerce templates nest Offer inside Product — match both to the buy box shoppers see.
Can I mark up aggregateRating I do not show on the page?
No. Rating and review fields must reflect visible, compliant review content. Fabricated ratings violate spam policies and mislead AI parsers as much as human readers.
Where should Organization schema live?
Most sites publish it on the homepage. If a canonical about page is your real brand URL, place it there instead — avoid duplicate, conflicting Organization blocks across templates.
Will FAQ schema bring back FAQ rich results in Google Search?
No. Google deprecated FAQ rich results from Search in May 2026; markup alone will not restore that display, though visible, accurate FAQs can still help AI parsing.
Can I add FAQPage schema without a visible FAQ section?
No. Markup has to mirror content a visitor can actually read on the same page. Hidden or mismatched FAQ data is misleading structured data, not a shortcut.
Should the homepage appear in BreadcrumbList?
Usually yes, at position 1, matching your visible breadcrumb UI and navigation pattern.
Is schema markup required for AI Overviews, ChatGPT or Perplexity citations?
Not documented as required. Google states structured data is not required for generative AI features in Search, but accurate markup still reduces the chance of a wrong quote.
Does accurate schema markup guarantee AI citations?
No. It only helps a parser read your facts correctly once it can reach your page. Whether a platform cites you still depends on crawl access, content quality and its own selection rules.
Related questions
- AI search visibility checklist for small businessPrioritized crawl, structure, and content steps in one list.
- How to optimize your website for AI searchFix crawl access, robots.txt, schema.org, and answer-first content.
- E-E-A-T signals for AI searchClear expertise, real experience, and trustworthy sources across key pages.
Updated
