Schema Markup for AI Search: What Helps and What Does Not

What schema markup is

• Definition: Schema markup (structured data) is code that labels the facts on your page, such as business name, address, article author or product price.
• Format: JSON-LD is the format Google recommends. It sits in a <script> tag and doesn't change how the page looks.
• Vocabulary: The labels come from Schema.org.
• Purpose: It helps machines read your page more reliably. It does not replace the visible content.
• Not a ranking switch: It's information, not a boost button.

What Google says

• For AI features: Google's guidance says optimizing for its AI features is still SEO, and no special markup, files or rewriting for AI is required.
• For rich results: Structured data can make a page eligible for specific rich results, but eligibility is not a guarantee.
• Match the page: Google's guidelines require markup to reflect the visible content.
• Not required for indexing: Pages can be indexed and shown without schema.
• Practical reading: Add schema where it's accurate and useful. Don't add it expecting AI visibility.

What the evidence shows

Image slot 2 · 1200x800 pxschema-ai-citations-evidence.webp

Alt text: Before and after bars showing little change in AI citations after schema was added to pages

AI image prompt

"A glass chart card on a dark navy background showing pairs of nearly equal before and after bars for three platforms, in cyan, blue and green, with a faint dotted trend line. Subtle grid, soft glow, minimal, no readable text, no logos."

Copy prompt

• The study: In May 2026, Ahrefs published a matched study of 1,885 pages that added JSON-LD between August 2025 and March 2026, compared with about 4,000 similar pages that did not.
• What it measured: Citations in Google AI Overviews, Google AI Mode and ChatGPT for 30 days before and after.
• Results: AI Mode and ChatGPT showed no change distinguishable from zero. AI Overviews showed a 4.6% decline, which was statistically significant, but the cause is unclear because both groups were already declining.
• The correlation: Ahrefs also found that AI-cited pages were almost three times more likely to have JSON-LD across 6 million URLs. The authors note that this can reflect correlation rather than cause.
• Its limits: The pages were already heavily cited, the window was 30 days, schema changes often came with other edits, and all schema types were pooled.
• Takeaway: Pages that get cited tend to have schema, but adding schema to a page didn't make it cited in this study.

Why schema still matters

• Clear facts: It states your business name, address, phone, hours and logo in a machine-readable way.
• Rich results: Some types still qualify for enhanced results in Google, such as product and review snippets, within Google's rules.
• Knowledge consistency: It reinforces the same details found on your Google Business Profile and directories.
• Search Console reports: Valid markup appears in enhancement reports, which makes problems easier to spot.
• Future-proofing: Other tools and platforms may read it even if the evidence for AI citations is weak now.
• Low cost: On most sites it's a one-time setup managed through a template or plugin.

Which schema types are worth adding

What has changed for FAQ and HowTo

• FAQ rich results: Google limited FAQ rich results in 2023 to well-known government and health sites. Most business sites no longer get the expandable FAQ in results.
• HowTo rich results: Google removed HowTo rich results in 2023.
• Should you remove them? Not necessarily. Visible FAQs still help readers. Markup that matches visible content is harmless, but don't expect a rich result.
• Rule: Only mark up questions and answers that appear on the page.
• Check before relying on any claim: Rich result rules change, so confirm in Google's current documentation.

A safe example: LocalBusiness

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Example Networks Pvt Ltd",
  "url": "https://www.example.com/",
  "logo": "https://www.example.com/logo.png",
  "telephone": "+91-00000-00000",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "12 Example Road",
    "addressLocality": "Mumbai",
    "addressRegion": "Maharashtra",
    "postalCode": "400001",
    "addressCountry": "IN"
  },
  "areaServed": "Mumbai",
  "openingHours": "Mo-Sa 10:00-18:00",
  "sameAs": ["https://www.linkedin.com/company/example"]
}

• Placeholders only: Replace every value with your real, visible details.
• Match everywhere: Use the same spelling and numbers as your page and Google Business Profile.
• Don't add what you don't have: Skip ratings or reviews unless you show real ones on the page.

How to implement it step by step

• Step 1: Audit what exists. Check your pages for existing markup from your theme or plugins, so you don't duplicate or conflict.
• Step 2: Choose types by page. Home: Organization and WebSite. Contact: LocalBusiness. Blog posts: BlogPosting and BreadcrumbList. Services: Service and BreadcrumbList.
• Step 3: Write JSON-LD. Use a template, plugin or hand-written script, matching visible content.
• Step 4: Add it in the page head or body. A <script type="application/ld+json"> tag works in either.
• Step 5: Validate. Test each template type, not just one page.
• Step 6: Deploy and monitor. Check Search Console enhancement reports after Google recrawls.
• Step 7: Keep it in sync. Update markup whenever your address, phone, hours or prices change.

How to check your markup

• Google's Rich Results Test: Shows which Google rich results a page is eligible for and flags errors.
• Schema Markup Validator: Checks general Schema.org syntax, not just Google features.
• Search Console: Enhancement reports show errors and warnings across the site.
• View source: Confirm the JSON-LD is in the HTML that crawlers receive, not only added after scripts run.
• Compare with the page: Every value in the markup should be visible somewhere on the page.

Mistakes that cause harm

• Markup that doesn't match the page: It can lead to manual actions or ignored markup.
• Fake or inflated ratings: Marking up reviews that don't exist is against Google's guidelines.
• Duplicate conflicting markup: A theme and a plugin both outputting different organisation details.
• Marking up hidden content: Questions or prices not visible to visitors.
• Outdated details: Old phone numbers or addresses in markup.
• Believing the hype: Paying for "AI schema" packages that promise citations.
• Skipping the content: Using markup to cover for thin pages.

Mumbai and India: local business markup

• Use LocalBusiness with a real address: Include addressCountry as IN and the correct PIN code.
• Describe your service area: Use areaServed for the areas you really cover.
• Match your Google Business Profile: Same name, category wording, phone and hours.
• Use international phone format: Include the +91 country code.
• Add languages only if real: Don't mark up languages you don't support.
• See also: Google Business Profile management keeps your listing in step with your site.

Limits and honest caveats

• One strong study, not the last word. Ahrefs's study is large and well designed, but it covers heavily cited pages, a 30-day window and pooled schema types.
• Other claims are weaker. Several articles claim large gains from schema without a control group. We haven't relied on them.
• Platforms change. A tool could start using structured data more directly, so keep your markup accurate.
• Rich result rules change. Check Google's documentation before relying on any feature.
• Our bias. We sell AI search optimization and SEO services and implement schema for clients, so we have an interest in explaining where it fits.

How our services help

• SEO services: Technical audits, including structured data checks.
• AI search optimization: Answer-ready pages, entity consistency and citation testing.
• Website development: Schema built into page templates.
• Google Business Profile management: Local details that match your markup.