Schema.org lists over 800 types of structured data. You don't need most of them. For a local business trying to get cited by AI tools like ChatGPT, Perplexity, and Google AI Overviews, five schema types do almost all the work: LocalBusiness, FAQPage, Article, AggregateRating, and BreadcrumbList. This article explains what each one does, why it affects AI citation rates, and exactly which fields to include.
Why structured data matters so much for AI search in 2026
AI tools don't rank pages the way Google's traditional algorithm does. They retrieve information — and the quality of that retrieval depends heavily on how clearly the information is labelled. A page with no schema is a box of unsorted documents. A page with structured data is a neatly labelled filing cabinet.
When ChatGPT is deciding between two local plumbers to recommend, all else being equal, it will cite the one whose website clearly states their business type, location, services, credentials, and what questions they can answer — in structured, machine-readable format. The other business might have all the same information, buried in paragraphs of flowing prose. But AI has to work to extract it, introducing the risk of misinterpretation or simply overlooking it.
Structured data doesn't make your business better. It makes your business legible to AI. That's worth a lot more than it sounds.
Type 1: LocalBusiness schema
LocalBusiness — tells AI who and where you are
LocalBusiness schema is the foundation. It identifies your business to AI tools by explicitly labelling your name, address, phone number, business type, opening hours, service area, and URL. Without it, every AI tool that encounters your website has to infer these facts from unstructured content — and inference introduces errors.
The most critical fields to include:
- @type — use the most specific subtype available: "Plumber", "Electrician", "HairSalon", "AutoRepair" rather than just "LocalBusiness"
- name — your exact trading name, matching your Google Business Profile
- address — PostalAddress with streetAddress, addressLocality, postalCode, addressCountry
- telephone — your primary phone number
- openingHoursSpecification — each day with opens/closes times
- areaServed — the geographic areas you cover
- url — your website URL
- priceRange — "£" to "££££" — helps AI match you to price-related queries
Type 2: FAQPage schema
FAQPage — pre-formats your answers for AI extraction
FAQPage schema labels your question-and-answer content in a format AI can extract directly. It's the single highest-impact schema type for AI citation rate — pages with FAQPage schema appear in 47% more Perplexity responses than those without. The reason is structural: AI tools are built to answer questions. FAQPage schema gives them pre-packaged answers that don't require interpretation.
The key rules:
- 4–8 questions is the optimal range — fewer doesn't provide enough coverage, more dilutes quality
- Every answer must start with the direct answer in the first sentence — not a preamble
- The Q&As in the schema must exactly match visible Q&A content on the page
- Keep answers to 2–4 sentences — concise and extractable
- Use real questions customers ask — check your enquiry emails, phone calls, and review content for source material
Type 3: Article schema
Article — signals expertise and freshness to AI
Article schema tells AI tools that a page contains authored content — not just a service listing. It includes the author's name, the publication date, and the headline. This is important for AI citation for two reasons: first, AI tools apply "expertise and authority" signals when choosing sources; second, the datePublished field tells AI the content is current — a major factor when AI tools are forming recommendations about anything time-sensitive.
Essential fields:
- headline — the article title, matching your H1
- datePublished — the original publication date in ISO 8601 format (e.g. "2026-06-22")
- dateModified — when you last updated it — keep this current
- author — your name and job title (Person type with name and jobTitle)
- publisher — your business name and logo URL
- description — a 1–2 sentence summary of the article content
A recent datePublished on a well-structured Article page is one of the key signals that differentiates a citable source from an outdated one in AI recommendations.
Type 4: AggregateRating schema
AggregateRating — tells AI your review score without them having to check
AggregateRating schema marks up your overall rating and review count directly on your website. When AI tools are comparing businesses to recommend, review scores are a significant factor — ChatGPT uses review data in 58% of local recommendations. AggregateRating schema makes this data immediately accessible without the AI needing to cross-reference external review platforms.
Fields to include:
- ratingValue — your average rating (e.g. "4.8")
- reviewCount — your total number of reviews
- bestRating — maximum possible rating ("5")
- worstRating — minimum possible rating ("1")
Important: AggregateRating must be nested inside your LocalBusiness schema, not added separately. And your rating must be real and current — Google cross-checks this data.
Type 5: BreadcrumbList schema
BreadcrumbList — helps AI understand your site structure
BreadcrumbList schema labels your page hierarchy — where a page sits within your website structure. It's less dramatic than LocalBusiness or FAQPage in terms of direct citation impact, but it's important for helping AI tools understand the relationship between your pages. A well-structured breadcrumb hierarchy signals topical organisation, which contributes to the "authority" signals AI uses when deciding whether to cite your site.
Fields to include:
- itemListElement — an array of ListItem, each with position, name, and item (URL)
- Example: Home › Services › Boiler Installation would be three ListItems at positions 1, 2, 3
BreadcrumbList also enables breadcrumb display in Google Search results, which improves click-through rates from traditional search — a secondary benefit on top of the AI visibility effect.
How to implement all five without a developer
For WordPress sites, a single plugin like Rank Math (free) covers LocalBusiness, Article, FAQPage, AggregateRating, and BreadcrumbList with no coding. It auto-generates most of the schema from your settings and page content.
For non-WordPress sites, generate the JSON-LD for each type manually or using Google's Structured Data Markup Helper, paste it into your page's <head> section, and verify with the Rich Results Test at search.google.com/test/rich-results.
If you're building a new site, use a platform that includes schema markup as standard. WeLaunchd builds LocalBusiness and FAQPage schema into every site as default — you don't need to configure anything.
If you can only add one schema type today: LocalBusiness. If you can add two: LocalBusiness + FAQPage. Together they cover the two most impactful citation signals for local business AI visibility. Add Article, AggregateRating, and BreadcrumbList as you build out your content — they compound the effect.
Get the complete AI Local Search guide series
Module 1 covers schema markup in full — with the exact LocalBusiness and FAQPage JSON-LD templates to copy into your site, explained step by step for non-technical business owners.
Structured data is only half the job — confirming it's actually improving your AI visibility is the other half. StackProven's review of Semrush's AI visibility toolkit covers one way to check.