Technical · 11 min read

The schema markup guide for AI search

FAQPage, Organization, Article, HowTo and LocalBusiness — what to deploy and why.

By the AEO GEO Marketing team · Updated

Key takeaways
  • Schema describes your content in a standard vocabulary.
  • Use JSON-LD with stable @id values and sameAs links.
  • Organization, Service, LocalBusiness, Article, Person and FAQPage cover most needs.
  • Only mark up content that is visible and accurate.
  • Validate before launch and monitor in Search Console.

Schema markup is structured data, usually JSON-LD, that tells machines exactly what a page is about. For AI search, it removes ambiguity: who you are, what you offer, where you operate and which questions you answer.

The essential types

  • Organization / ProfessionalService: your name, logo, contact details, profiles (sameAs) and area served.
  • LocalBusiness: address, opening hours and geo-coordinates for each location or city page.
  • FAQPage: explicit question-and-answer pairs, perfect for AEO.
  • Article: headline, author, dates and publisher for editorial content.
  • HowTo: step-by-step instructions.
  • Service / Product / Offer: what you sell and for how much.
  • BreadcrumbList: site hierarchy.

Rules that matter

  • Mark up only what is visible on the page.
  • Keep facts consistent with your other profiles and directories.
  • Validate every template with the Schema.org validator and Google's Rich Results Test.
  • Connect entities with @id references so your Organization, people and services form one graph.

Schema and AI citations

Schema alone will not get you cited. It does make your content easier to understand, extract and trust. Combined with answer-first content and authority signals, it is one of the highest-leverage technical moves in GEO.

What is schema markup and how does it work?

Schema markup is structured data added to a web page to describe what the page is about in a standard vocabulary defined by Schema.org. Instead of an engine guessing that a block of text is a business address or a list of questions, the markup states it explicitly. Google, Bing and other engines use this information to understand content, display rich results and connect pages to entities in their knowledge graphs.

The recommended format is JSON-LD: a script block in the page’s HTML that contains the structured data as JSON. It sits separately from the visible content, which makes it easy to add and maintain.

Which schema types should every business use?

Organization

Describes your company: name, logo, URL, contact details and links to official profiles (sameAs). Place it on the home page or site-wide.

WebSite

Describes the site itself and can declare its name and search function.

LocalBusiness or ProfessionalService

For businesses that serve specific areas. Include address where relevant, areaServed, opening hours and price range.

Service and Offer

Describe what you sell and, where appropriate, prices and currencies.

BreadcrumbList

Shows where each page sits in the site hierarchy.

Article and Person

For blog posts and guides, with headline, author, publisher, datePublished and dateModified. Link authors to Person profiles with credentials.

FAQPage

For genuine lists of questions and answers visible on the page.

AI engines rely on understanding entities and facts correctly. Schema helps in three ways:

  1. Disambiguation. It tells engines exactly which organisation, person or place a page refers to.
  2. Fact extraction. Prices, locations, dates and answers are presented in a predictable format.
  3. Entity connections. sameAs links and consistent identifiers connect your website to profiles elsewhere, strengthening your entity in knowledge graphs.

Schema is not a ranking shortcut, and engines may ignore markup that does not match visible content. Used accurately, it reduces ambiguity, which is a precondition for being trusted and cited.

A practical JSON-LD example

A minimal Organization block includes the organisation’s type, name, URL, logo, contact email and a list of official profiles under sameAs. A FAQPage block includes a mainEntity list, where each item is a Question with a name and an acceptedAnswer containing the answer text.

Best practice is to give important entities a stable @id (for example, your home page URL followed by #organization) and reference that @id from other blocks, such as the publisher of an Article or the provider of a Service. This creates a connected graph rather than isolated fragments.

How do you validate schema?

  • Use Google’s Rich Results Test to check eligibility for rich results and spot errors.
  • Use the Schema Markup Validator at validator.schema.org to check general Schema.org syntax.
  • Monitor the Enhancements reports in Google Search Console for errors across the site.
  • Re-test after template changes, because a single bug can break markup on thousands of pages.

Common schema mistakes

  • Marking up content that is not visible on the page.
  • Using FAQPage for promotional copy rather than genuine questions and answers.
  • Inconsistent business names, addresses or phone numbers between schema and the page.
  • Leaving out dateModified on articles that are updated.
  • Duplicating conflicting Organization blocks across templates.
  • Forgetting to update prices and offers when they change.

Schema for multi-location businesses

If you serve many cities, give each location page its own LocalBusiness or ProfessionalService block with the correct areaServed, and link it back to the main Organization with parentOrganization or a shared @id. Combine this with unique, useful local content on each page, because markup alone does not make a page valuable.

Which properties matter most for each type?

Organization

name, url, logo, description, email, telephone, address (if public), sameAs (official profiles) and contactPoint.

LocalBusiness / ProfessionalService

All Organization properties plus areaServed, openingHoursSpecification, priceRange and geo coordinates where you have a physical location.

Service

name, description, provider (linked to your Organization), areaServed and offers with price and priceCurrency where appropriate.

Article / BlogPosting

headline, description, image, author (Person or Organization), publisher, datePublished, dateModified and mainEntityOfPage.

Person

name, jobTitle, worksFor, image, sameAs (LinkedIn and other profiles), knowsAbout and alumniOf where relevant.

FAQPage

mainEntity, with each Question having a name and an acceptedAnswer containing the full visible answer text.

How do you implement schema at scale?

For sites with many pages, schema should be generated by templates rather than written by hand. Store key facts, such as company details, service prices and author profiles, in one place and render them into JSON-LD automatically. This keeps markup consistent and makes updates easy. After deployment, monitor Search Console and run periodic crawls that validate markup across all templates.

Does schema help with ChatGPT and Perplexity?

AI search engines do not all publish how they use structured data. However, schema makes facts explicit and consistent, helps search indexes that these engines rely on understand your pages, and strengthens entity data in knowledge graphs. Even where an engine reads only the visible text, the same discipline that produces good schema, clear facts stated consistently, also produces content that is easier to cite.

How do you maintain schema over time?

Treat schema as part of your content, not a one-off project. Update dateModified when content changes, keep prices and offers current, add new authors and services as they appear, and remove markup for things that no longer exist. Assign an owner and review schema whenever templates change.

Where should JSON-LD go on the page?

JSON-LD can be placed in the head or body of the HTML. Most sites put it in the head so it loads early and is easy to find. What matters most is that it is present in the HTML the crawler receives. If your site injects schema with JavaScript, test with Google’s URL Inspection tool to confirm that the rendered page includes it.

How do @id and sameAs work?

An @id is a unique identifier for an entity within your structured data, usually a URL with a fragment such as https://example.com/#organization. Using the same @id wherever the entity appears tells engines that every reference is the same thing. sameAs lists other URLs that represent the same entity, such as your LinkedIn page, Wikipedia entry or Crunchbase profile. Together they help engines consolidate everything known about your brand into one entity.

Schema for AI search: what is changing?

Search engines periodically change which schema types produce visible rich results. For example, Google has reduced how often FAQ and HowTo rich results are shown for most sites. That does not make the markup useless: it still helps engines understand content, and AI systems benefit from clear, structured facts. Focus on accurate markup that describes your content honestly, rather than chasing specific visual features.

Example: schema for a service business

A service business might use the following on its main pages:

  • Home page: Organization (with @id, logo, contact details and sameAs) and WebSite.
  • Service page: Service linked to the Organization as provider, with areaServed and Offer.
  • Location page: LocalBusiness or ProfessionalService with address or areaServed, opening hours and a link to the parent Organization.
  • Team page: Person entries for key experts, linked to the Organization with worksFor.
  • Blog posts: BlogPosting with author, dates and publisher.
  • All pages: BreadcrumbList.

How to test schema before launch

Before launching new templates, test sample pages in the Rich Results Test and the Schema Markup Validator, check that every value matches the visible page, and confirm that IDs are consistent across templates. After launch, watch Search Console enhancement reports for errors and spot-check pages after any template or CMS update.

Schema for e-commerce sites

Online stores should use Product with name, image, description, brand, sku and gtin where available, plus Offer with price, priceCurrency, availability and shipping details, and AggregateRating and Review where genuine reviews are shown. Keep markup synchronised with your product feed so prices and availability are always correct; mismatches can cause rich results to be removed and confuse AI shopping answers.

Frequently asked questions about schema

Does schema improve rankings?

Schema is not a direct ranking factor, but it can make pages eligible for rich results and helps engines understand content, which supports visibility.

Can I use more than one schema type on a page?

Yes. Many pages combine Organization, WebPage, BreadcrumbList and a main type such as Article or Service. Link them with @id references.

Should every page have FAQPage schema?

No. Use it only where the page contains genuine questions and answers that users can see.

Schema for professional services firms

Law firms, accountants, consultants and agencies benefit most from Organization or ProfessionalService for the firm, Person for each partner or expert with credentials and sameAs links, Service for each practice area, and FAQPage for genuine client questions. Linking experts to the firm and to their articles helps engines connect expertise to your brand, which is especially important for YMYL topics such as legal and financial advice.

A quick schema audit you can do today

Open your home page, a service page and a blog post in Google’s Rich Results Test. Check that each page has at least Organization (or a reference to it), BreadcrumbList and the right main type. Confirm that names, contact details and prices match the visible page. Note any errors or warnings, and fix template-level issues first, because they affect every page built from that template.

Schema checklist

  • Organization and WebSite site-wide, with a stable @id.
  • BreadcrumbList on every page below the home page.
  • Service, Offer or Product on commercial pages.
  • LocalBusiness or ProfessionalService on location pages.
  • Article with author Person, datePublished and dateModified on content.
  • FAQPage only where real questions and answers are visible.
  • Validate, monitor and update regularly.

Continue with these guides:

Or see how we put this into practice:

Want this done for you?

Get a free 48-hour GEO audit that shows exactly where AI engines cite you — and where they cite your competitors.