The schema types that matter most for AI search are FAQPage (350% citation increase), Organisation with sameAs (entity verification), Article with Author/Person (3x citation lift), Speakable (127% increase but under 10% adoption), HowTo, and BreadcrumbList. 71% of pages cited by ChatGPT use structured data. The priority is not having schema — it is having the right schema.

Schema priority ranking for AI search

Priority	Schema type	AI impact	Current adoption
1	FAQPage	350% citation increase	Moderate
2	Organisation	Entity verification	56% have it
3	Article + Author (Person)	3x citation lift	4% use Person
4	Speakable	127% citation increase	Under 10%
5	HowTo	Step extraction for AI answers	Low
6	BreadcrumbList	Site structure signal	Moderate

FAQPage: the highest-impact type

FAQPage schema delivers a 350% increase in AI citation rates — the largest impact of any single schema type. It works because AI platforms are specifically looking for question-and-answer pairs, and FAQPage markup tells them exactly where those pairs are in your content. Yet our audit found 52% of sites with FAQ content on the page do not use FAQPage schema.

The reason FAQPage is so effective is structural: AI platforms break queries into sub-questions, and FAQPage schema provides pre-structured answers that map directly to those sub-questions. It reduces the effort AI needs to extract a citable answer from your page.

Organisation schema: the foundation

Organisation schema is the identity layer. It tells AI platforms your business name, type, location, contact details, and — critically — your sameAs links to LinkedIn, Google Business Profile, Crunchbase, and other platforms where your entity exists.

AI platforms cross-reference these signals. If your Organisation schema says you are a digital marketing agency in London, but your Google Business Profile says "SEO consultant" and your LinkedIn says "technology company", the conflicting signals reduce AI confidence in your entity.

Our audit found 44% of UK SME websites lack Organisation schema entirely.

Article + Author (Person): the credibility multiplier

Content with Article schema that includes a named author with Person schema is 3x more likely to be cited by AI platforms. This is because AI systems assess content credibility, and a named author with verifiable credentials is a strong trust signal.

Person schema connects the author to their LinkedIn profile, publications, and professional credentials via sameAs links — creating a verifiable identity that AI can cross-reference.

Only 4% of UK businesses currently use Person schema — making it one of the biggest competitive gaps in AI search.

Speakable: the overlooked opportunity

Speakable schema marks specific sections of content as suitable for text-to-speech and direct extraction. It delivers a 127% citation increase but is used by fewer than 10% of websites. This is the single biggest gap between impact and adoption in AI schema.

Speakable tells AI platforms: "This paragraph is the extractable answer." It is essentially a spotlight on your most citable content. See our full analysis of the Speakable adoption gap.

HowTo and BreadcrumbList

HowTo schema structures step-by-step processes in a format AI can directly extract and present. If your content includes instructions, tutorials, or guides, HowTo schema makes those steps machine-readable.

BreadcrumbList signals your site's hierarchical structure, helping AI platforms understand content relationships and topic authority. It is a supporting signal rather than a primary driver, but it contributes to the overall structured data picture.

The 71% statistic

Research shows that 71% of pages cited by ChatGPT use some form of schema markup. This does not mean schema alone guarantees citation — but it does mean that pages without schema are significantly underrepresented in AI answers. Schema is a necessary condition, not a sufficient one.

Read our complete schema markup guide for AI search · Get a free schema audit

What Schema Markup Matters Most for AI Search?