Guide

    Schema markup for AI search: what to add, and what not to bother with

    Structured data tells search engines and AI tools exactly what your business is. The schema types worth adding, the ones Google retired, and common mistakes.

    By NoorRank · Updated

    In short

    Schema markup is structured data, usually JSON-LD, that states facts about your business in a format machines read without guessing. Google says it isn't required for AI Overviews, but it helps search engines understand who you are, what you offer and where. Start with Organization, your business type, your services and your key pages, and keep every fact consistent with what the page says.

    What schema markup is

    Schema markup uses the shared vocabulary at schema.org to label facts on a page: this is an organisation, this is its address, these are its opening hours, this page is an article written on this date. Most sites add it as a block of JSON-LD in the page's HTML, invisible to visitors.

    To a person, your contact page obviously shows your address. To a machine, it is a string of text that might be an address. Schema removes the guesswork.

    Is it required for AI search?

    No. Google says structured data isn't required to appear in AI Overviews or AI Mode. But it does help Google understand your content, can make pages eligible for rich results, and gives every system that reads your page an unambiguous statement of your facts. For businesses with common names, several locations or bilingual sites, that clarity matters a lot.

    The types worth adding first

    Schema typeUse it forKey properties
    OrganizationYour business as a whole, on the homepagename, url, logo, description, sameAs (your profiles), contactPoint
    LocalBusiness or a subtypeEach physical location (MedicalClinic, Dentist, RealEstateAgent and so on)address, geo, openingHoursSpecification, telephone, areaServed
    ServiceEach service pagename, serviceType, provider, areaServed, description
    Product and OfferAnything with a pricename, price, priceCurrency, availability
    Article or BlogPostingGuides and newsheadline, datePublished, dateModified, author
    BreadcrumbListEvery page below the homepageThe path from home to the page
    FAQPagePages with real question-and-answer contentQuestion and acceptedAnswer pairs

    Connect it with @id

    The most overlooked step: give your organisation one stable identifier, such as https://yoursite.com/#organization, and reference it from every page's markup (as the publisher of articles, the provider of services, the parent of each location). This turns separate blocks of markup into one connected picture of your business, which is exactly the entity view AI systems work from.

    Use sameAs to link that entity to your official profiles: LinkedIn, Google Business Profile, Wikidata if you have an entry, and major directories. It tells machines that these are all the same business.

    What Google retired

    • FAQ rich results have been limited since 2023 to well-known government and health websites. FAQPage markup is still valid and still describes your content, but most sites no longer get the expandable FAQ display in results.
    • HowTo rich results were removed. The markup does no harm, but don't expect a visual result from it.

    Common mistakes

    • Marking up things the page doesn't show. Structured data must match visible content. Invented ratings or hidden FAQs can lead to a manual action.
    • Self-serving review markup. Google doesn't allow review or rating markup about your own business on your own site.
    • Contradictions. A founding year, phone number or address in your schema that differs from your page text, or from your profiles elsewhere.
    • Generic types. Using plain LocalBusiness when a specific subtype exists, such as MedicalClinic or Restaurant.
    • Never testing. Run key pages through Google's Rich Results Test and the Schema Markup Validator after every change.

    Questions people ask

    Does schema markup improve rankings?

    Not directly. It helps search engines understand your pages and can make them eligible for rich results, which can improve clicks. Its bigger value for AI search is clarity about your facts.

    Should I use JSON-LD or microdata?

    JSON-LD. It is Google's recommended format and sits in one block, separate from your page layout, which makes it easier to maintain.

    Is FAQPage schema still worth adding?

    If the page has genuine questions and answers, yes, because it describes the content clearly. Just don't expect the FAQ rich result, which Google now shows only for certain government and health sites.

    See what AI says about your business

    The free Ghost Audit shows how ChatGPT, Perplexity, Gemini and Google's AI Overviews describe you today, and who they recommend instead.