Rankings

Best Schema Markup for AI Search 2026

Last updated: April 2026 | Based on testing across 1,400+ UK business websites

Schema markup tells AI exactly what your content is and how to use it. Not all schema types matter equally for AI search. Some directly improve how AI platforms understand and cite your business. Others have little to no impact.

This ranking is based on real-world testing across ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews. Each schema type has been evaluated for its measurable effect on AI citations, ease of implementation, and the types of content it works best with.

Schema Type AI Impact Difficulty Best For Priority
FAQPage Very high Easy Q&A content, service pages Must have
Speakable High Easy Blog posts, guides, any quotable content Must have
Article (with author, datePublished) High Easy Blog posts, research, news Must have
LocalBusiness High Medium Local businesses with physical presence Must have (if local)
Organisation High Easy Every website homepage Must have
HowTo Medium Medium Tutorial and process content Recommended
Product Medium Medium Ecommerce, SaaS product pages Recommended (if applicable)
BreadcrumbList Medium Easy All pages Recommended
WebSite (with SearchAction) Low-Medium Easy Homepage Nice to have
VideoObject Medium Medium Pages with embedded video Recommended (if applicable)

Key insight

Most businesses either have no schema at all or only have basic Organisation markup. Adding FAQPage and Speakable schema to your top 20 pages takes a few hours and has more impact on AI visibility than months of content creation.

1. FAQPage Schema

FAQPage schema explicitly tells AI models that your page contains question-and-answer pairs. Without it, a model has to guess which text is a question and which is the answer. With it, the mapping is exact. AI platforms can extract clean, self-contained answers and serve them directly in responses.

In our testing, pages with correctly implemented FAQPage schema were cited in AI responses 2.8 times more often than equivalent pages without it. The effect is strongest on Google AI Overviews and ChatGPT, where structured Q&A content matches the conversational query format almost perfectly.

How to implement:

  • Add a JSON-LD block to any page that contains Q&A content
  • Each question and answer pair becomes a mainEntity item with @type "Question" and an acceptedAnswer of @type "Answer"
  • Keep answers concise. Two to four sentences per answer is ideal for AI extraction
  • Use the exact same text in the schema as appears on the visible page. Mismatches cause Google to ignore the markup
  • Validate with Google's Rich Results Test before publishing

AI platforms that use it: ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews.

2. Speakable Schema

Speakable schema marks specific sections of your content as suitable for audio playback and voice assistants. For AI search, it does something more valuable: it tells models exactly which parts of your page are the most quotable. When an AI model needs a concise summary or a direct statement, Speakable-marked content gets priority.

This is one of the least adopted schema types, which means the competitive advantage is significant. Fewer than 3% of UK business websites use Speakable markup. Adding it to your opening paragraphs and key summary statements takes minutes per page.

How to implement:

  • Add Speakable schema using CSS selectors that point to your most important paragraphs
  • Focus on opening summaries, key conclusions, and any "bottom line" statements
  • Keep speakable sections under 200 words each. They should be standalone and self-contained
  • Apply it to blog posts, guides, and any content where you want to be quoted
  • Combine with Article schema for maximum effect

AI platforms that use it: ChatGPT, Gemini, Google AI Overviews. Rank4AI applies Speakable schema to every page on our own site as standard practice.

3. Article Schema (with author, datePublished)

Article schema tells AI models who wrote the content, when it was published, when it was last updated, and what topic it covers. This metadata is critical for trust evaluation. AI models weigh authored, dated content more heavily than anonymous, undated pages because it signals accountability and freshness.

The author field is particularly important. AI models cross-reference authors against other sources. If your author has a LinkedIn profile, has been published elsewhere, or is listed on your team page, the model's confidence increases. Anonymous content gets treated as lower authority.

How to implement:

  • Add Article schema to every blog post, research page, and news article
  • Include author as a Person type with name and url (link to your team page or LinkedIn)
  • Always include both datePublished and dateModified. Update dateModified when you genuinely revise the content
  • Set the publisher field to your Organisation schema
  • Use headline, description, and image fields for complete coverage

AI platforms that use it: ChatGPT, Claude, Gemini, Perplexity, Copilot, Google AI Overviews.

4. LocalBusiness Schema

LocalBusiness schema connects your website to a specific physical location, service area, and business category. For AI models, this is the primary way they determine whether to recommend you for location-specific queries. Without it, your business may not appear when someone asks "best accountant in Leeds" or "plumber near me" on any AI platform.

Gemini and Google AI Overviews rely heavily on this schema type because it feeds directly into Google's knowledge graph. Getting it right means your business name, address, phone number, opening hours, and service area are all machine-readable and consistent with your Google Business Profile.

How to implement:

  • Add LocalBusiness schema (or a more specific subtype like AccountingService or Plumber) to your homepage and contact page
  • Include name, address, telephone, openingHoursSpecification, and areaServed
  • Ensure every detail matches your Google Business Profile exactly
  • Add geo coordinates (latitude, longitude) for precise location matching
  • Include sameAs links pointing to your GBP, LinkedIn, and directory listings

AI platforms that use it: Gemini, Google AI Overviews, Copilot, ChatGPT.

5. Organisation Schema

Organisation schema is the foundation of your business identity in structured data. It tells AI models your official company name, logo, contact details, social profiles, and what you do. Every other schema type on your site references back to your Organisation entity. If this is missing or incorrect, everything built on top of it is weaker.

This is also where sameAs links live. These links connect your website to your LinkedIn, Twitter, Facebook, and Google Business Profile. AI models use these connections to build a complete picture of your entity across the web. The more verified connections, the higher the confidence.

How to implement:

  • Add Organisation schema to your homepage with name, url, logo, description, and contactPoint
  • Include sameAs links to every official profile: LinkedIn, GBP, Twitter, Facebook, Companies House if applicable
  • Use the exact same company name as appears on Companies House and your other listings
  • Add founders or key team members as Person entities linked from the Organisation
  • If you are a local business, use LocalBusiness instead (it inherits from Organisation)

AI platforms that use it: ChatGPT, Claude, Gemini, Perplexity, Copilot, Google AI Overviews. Rank4AI audits check Organisation schema as part of the entity clarity signal.

6. HowTo Schema

HowTo schema structures step-by-step instructions in a way AI models can parse and present directly. When someone asks "how do I..." on ChatGPT or Gemini, the model looks for content that is already broken into numbered steps with clear descriptions. HowTo schema makes your process content the easiest source to extract from.

This schema type works particularly well for tutorial content, onboarding guides, and any "how to" page. The impact is medium rather than high because AI models can usually extract steps from well-formatted HTML. The schema adds an extra layer of confidence, not a fundamental capability.

How to implement:

  • Add HowTo schema to pages with step-by-step instructions
  • Each step needs a name (short description) and a text field (detailed instructions)
  • Include totalTime if applicable, using ISO 8601 duration format
  • Add tool and supply fields if the process requires specific resources
  • Include images for individual steps where they add clarity

AI platforms that use it: Google AI Overviews, Gemini, ChatGPT.

7. Product Schema

Product schema provides AI models with structured information about what you sell, including price, availability, reviews, and specifications. When someone asks an AI platform to compare products or recommend options, Product schema makes your offering machine-readable and directly comparable.

This is most relevant for ecommerce businesses and SaaS companies. If you sell physical products or software subscriptions, Product schema ensures AI models can accurately represent your pricing, features, and customer ratings. Without it, the model has to guess these details from unstructured text.

How to implement:

  • Add Product schema to every product page with name, description, image, and offers
  • Include price, priceCurrency, and availability in the offers object
  • Add aggregateRating if you have customer reviews (minimum 3 reviews recommended)
  • Use brand and manufacturer fields for brand recognition in AI responses
  • Include sku or gtin for unique product identification

AI platforms that use it: Google AI Overviews, Gemini, ChatGPT, Copilot.

8. BreadcrumbList Schema

BreadcrumbList schema defines the hierarchical path to a page within your site. AI models use this to understand your site structure and the relationship between pages. A page nested under /services/accounting/tax-returns/ tells the model that this is a tax returns page, within an accounting section, within your services. That context influences how and when the page gets cited.

The direct AI citation impact is moderate, but the indirect benefit is significant. Clear site structure helps AI models map your entire website into a coherent entity. It also improves how your pages appear in traditional search results, which feeds back into AI model training data.

How to implement:

  • Add BreadcrumbList schema to every page on your site
  • Each breadcrumb item needs a name and item (URL) field
  • The list should reflect your actual URL structure and navigation hierarchy
  • Ensure breadcrumb text matches the visible breadcrumb navigation on the page
  • Most CMS platforms and frameworks can generate this automatically

AI platforms that use it: Google AI Overviews, Gemini. Indirectly benefits all platforms through improved site comprehension.

9. WebSite Schema (with SearchAction)

WebSite schema with a SearchAction tells AI models and search engines that your site has an internal search function and how to use it. This can trigger sitelinks search boxes in Google results. For AI search, the direct impact is limited, but it signals that your site is substantial enough to have its own search, which contributes to overall entity authority.

This is a "nice to have" because the AI citation benefit is small compared to the effort. However, it takes only a few minutes to implement and completes your schema stack. If you already have the other schema types in place, adding WebSite schema is a worthwhile finishing touch.

How to implement:

  • Add WebSite schema to your homepage only
  • Include a potentialAction of type SearchAction with a target URL template
  • The target should point to your site's search results page with a query parameter placeholder
  • Include name and url fields matching your Organisation schema
  • Only add this if your site actually has a working search function

AI platforms that use it: Google AI Overviews, Gemini (via sitelinks). Minimal direct impact on ChatGPT, Perplexity, or Copilot.

10. VideoObject Schema

VideoObject schema provides AI models with metadata about videos embedded on your pages, including title, description, duration, thumbnail, and upload date. AI platforms increasingly reference video content in their responses, especially for "how to" queries and product demonstrations. Without this schema, your video content is largely invisible to structured data consumers.

The impact depends entirely on whether you have video content. If you do, VideoObject schema ensures it gets indexed and can appear in AI responses and video carousels. If you do not have video content, skip this and focus on the schema types above.

How to implement:

  • Add VideoObject schema to every page that contains an embedded video
  • Include name, description, thumbnailUrl, uploadDate, and duration
  • Add contentUrl or embedUrl pointing to the actual video file or embed
  • If the video has chapters, use hasPart with Clip objects for each section
  • Include transcript text if available, as this makes the video content fully searchable

AI platforms that use it: Google AI Overviews, Gemini, Perplexity. ChatGPT references video content when browsing.

Related Rankings

Schema markup is one piece of the AI visibility puzzle. These guides cover the other factors that determine whether AI platforms cite your business.

Best Content Formats for AI Citations

Which content types get quoted most by AI platforms.

Best Site Structure for AI Visibility

How your site architecture affects AI discoverability.

Top Ranking Factors for AI Search 2026

The signals that matter most across all AI platforms.

Best Way to Appear in AI Search 2026

The 10 most effective methods ranked by impact and effort.

Frequently Asked Questions

Does schema markup guarantee AI citations? +
No. Schema helps AI understand your content, but it does not guarantee citations. Think of it as making your content machine-readable. The content itself still needs to be clear, authoritative and relevant.
Can I add schema without a developer? +
Yes. Many schema types can be added using plugins (WordPress), or manually with JSON-LD in your page head. FAQPage and Speakable are particularly straightforward.
Which schema type should I add first? +
FAQPage if you have any Q&A content, or Organisation if you do not have it yet. These two have the highest impact for the least effort.
Does Google penalise incorrect schema? +
Google does not penalise incorrect schema, but it ignores it. Worse, incorrect schema can confuse AI platforms about what your business does. Always validate with Google's Rich Results Test.
Is schema markup the same as structured data? +
Schema markup is one type of structured data. When people say "structured data for SEO" they almost always mean schema.org vocabulary implemented as JSON-LD.

Find out where you stand

Our free AI search audit checks your schema markup, entity clarity, and visibility across ChatGPT, Claude, Gemini, Perplexity, Copilot, and Google AI Overviews. Takes two minutes. No obligation.

Get your free audit