Search engines and AI engines are both trying hard to "read" web pages, but machines have no semantic intuition — they rely on structured clues. Structured data (Schema markup) is a page description written for machines: using the shared vocabulary of schema.org, it marks up "this is a company, this is an article, these are FAQs" so that Baidu, Google and even AI engines like Doubao and DeepSeek can accurately understand what a page is about. For any business site doing technical SEO, this is the lowest-cost, highest-certainty foundational work available.
On the payoff side, there are verifiable public studies. AIOSEO's long-running SERP feature tracking shows rich results capturing roughly 58% of clicks, against 41% for non-rich results. Google's own published case studies point the same way: Rotten Tomatoes saw a 25% higher click-through rate (CTR) on pages with structured data, and Nestlé reported an 82% higher CTR for pages appearing as rich results. Google officially states that structured data is not a direct ranking factor, but it unlocks rich results — star ratings, FAQ, breadcrumbs — that indirectly bring greater visibility and clicks. For AI search, structured data lets LLMs extract brand facts at lower cost, which is exactly why it matters just as much in GEO optimization.
What is structured data: a page description for machines
schema.org is the shared structured-data vocabulary maintained by Google, Microsoft, Yahoo and Yandex. Three encoding formats exist: JSON-LD (recommended), Microdata and RDFa. JSON-LD is injected as a script tag, does not affect rendering, and is the cheapest to maintain — it is Google's officially recommended format, and Baidu supports it as well.
A common misunderstanding is that structured data is "SEO code you copy once and forget". In practice it is a live layer of your page: every time the content changes — a new service, a new contact number, a new article — the markup should change with it. A mismatch between what the page says and what the schema declares is one of the most frequent causes of invalid or ignored markup.
Validate every page's markup with a testing tool before launch: invalid JSON-LD is silently ignored by crawlers and only adds noise. For bilingual sites, localize schema strings so the JSON-LD on your Chinese pages carries Chinese values, while keeping entity identifiers (the company name in both languages) consistent so engines match them to a single entity.
A minimal Article-type JSON-LD example:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "SEO vs GEO: The New Traffic Landscape of Search and AI",
"datePublished": "2026-08-04",
"author": { "@type": "Organization", "name": "Shanghai Zheming Information Technology Co., Ltd." }
}
</script>
Six Schema types every business site should use
| Schema type | What it does | Possible search enhancement |
|---|---|---|
| Organization | Declares company name, address, contact, brand info | Knowledge graph display, entity recognition |
| WebSite + SearchAction | Declares site identity and on-site search | Sitelinks in search results |
| BreadcrumbList | Marks the page's breadcrumb hierarchy | Breadcrumb path in search results |
| Article / NewsArticle | Marks article title, author, publish date | Rich snippets, top-story images |
| FAQPage | Marks frequently asked questions and answers | Structured Q&A that RAG crawlers can extract |
| Product / Service | Declares offerings, prices, features | Product-rich results, AI service entity matching |
The first five types cover the classic SEO base; the sixth (Product/Service) has grown in importance because AI engines match "service capability" to "brand entity" when answering recommendation-style questions. If your site sells or serves multiple lines, one Service or Product block per offering keeps the entity graph clean.
Structured data for Baidu, Google and AI search
The first layer of value is in the Baidu ecosystem. According to Baidu Search Resource Platform documentation, Baidu supports structured data markup; properly submitted markup helps the engine understand content more accurately, and some industries can obtain enhanced display such as rich snippets — a meaningful bonus for Chinese-language search optimization.
The second layer is in the Google ecosystem. According to Google Search Central documentation, correct structured data can qualify pages for Rich Results — breadcrumbs, FAQs, ratings and more — which directly lift click-through rates in search results. Note the May 2026 change: Google deprecated the FAQ rich result in search results display, but the FAQPage schema type itself remains valid and is still crawled by Bingbot, PerplexityBot and other RAG-oriented crawlers. In other words, the markup's value for AI inclusion outlived its display value on Google.
The third layer is in AI search. Large language models' retrieval-augmented generation (RAG) pipelines depend heavily on parseable entity information: Organization, Service and FAQPage markup helps AI engines accurately associate your brand entity with your service capabilities. In other words, structured data is not only the technical foundation of SEO, but also a prerequisite for GEO optimization — AI engines must be able to "read" your content before they will cite it.
Implementation and validation: avoiding the silent-failure trap
- Audit what you already have. Run existing pages through Google's Rich Results Test and Baidu's structured data testing tool to find broken or outdated markup.
- Deploy JSON-LD per page. Put Organization markup site-wide, Article markup on news pages, FAQPage on Q&A pages, and Service/Product markup on offering pages.
- Keep schema in sync with content. Update datePublished when you revise an article; change contact and address values in Organization markup the day they change.
- Monitor in Search Console. The Structured Data report shows which pages have valid markup and which have errors; Baidu's search resource platform offers equivalent diagnostics.
- Localize for bilingual sites. Use translated strings in Chinese-page markup while keeping entity identity consistent across languages.
A typical failure scenario: a company copies Article schema onto every page including contact pages, forgets to update the date, and then wonders why no rich result appears. The markup is not broken — it is simply irrelevant to the page. Keeping types matched to page purpose is the single most important discipline.
If you want AI engines to cite your brand in answers, structured data is the first technical step. Shanghai Zheming Information Technology Co., Ltd. builds Schema deployment and validation (Google Rich Results Test, Baidu structured data testing tools) into its website development and SEO optimization services; pricing is subject to our quotation. Call +86 18917757529 or email jaysun@widesight.cn for a consultation. For more on corporate site essentials, see B2B Website Guide: Build a Site That Generates Leads.
FAQ
Q1: Does structured data directly improve rankings?
Google states that structured data is not a direct ranking factor; it works by unlocking rich results and improving how content is understood, which indirectly lifts CTR and visibility. Treat it as an enabler rather than a lever.
Q2: Is FAQPage schema still worth deploying after the May 2026 change?
Yes. Google deprecated the FAQ rich results display in May 2026, but FAQPage remains a valid schema.org type that Bingbot, PerplexityBot and RAG crawlers still read. For AI search and GEO purposes, structured Q&A is more valuable now than it was for display.
Q3: JSON-LD, Microdata or RDFa — which should we choose?
JSON-LD. It is Google's recommended format, does not alter page rendering, and is the easiest to maintain and validate. Baidu supports structured data as well, and JSON-LD works across both ecosystems.
Q4: Should we localize schema strings on a bilingual site?
Yes. Chinese pages should carry Chinese values in their JSON-LD, and entity identifiers (for example the company name in both languages) must stay consistent so engines recognize one entity, not two.
Q5: How is structured data related to AI search and GEO?
RAG pipelines extract entity information from parseable markup; Organization, Service and FAQPage help AI engines match your brand to its capabilities. Structured data is the technical foundation for both SEO and GEO.
Related reading
This article was written by Zheming Digital Communication Research Institute. Data updated to 2026. Sources: AIOSEO SERP feature tracking (2026), Google case studies on structured data (Rotten Tomatoes, Nestlé), Google Search Central documentation (structured data gallery updated 2026-06-23), Baidu Search Resource Platform documentation. Consultation: +86 18917757529 ｜ jaysun@widesight.cn.