Does schema markup help you get cited by ChatGPT or Google AI Overviews?

Not directly. Google states that its AI features need no special markup, and other assistants mostly read the visible text of pages they retrieve. Structured data helps search engines understand your business accurately, which supports everything built on search, but it does not earn citations on its own.

Which schema types should a local business use?

Organization markup for the business as a whole, a specific LocalBusiness subtype for each location with address, phone and opening hours, Article markup for editorial content and BreadcrumbList for site structure. Add Product, Event or Service markup only where those things are genuinely on the page.

Is FAQ schema still worth adding?

For most businesses it no longer produces a visible search result, because Google limited FAQ rich results to authoritative government and health sites. It does no harm if it matches a visible FAQ on the page, but it is not a strategy for search or AI visibility.

Do I need a developer to add structured data?

Not always. Platforms such as WordPress handle basic markup through SEO plugins, though the output should be checked. Custom sites, multi-location organizations and bilingual sites benefit from a developer generating markup from the same data that builds the page, so the two never drift apart.

Structured Data, Schema Markup and AI Citations: What Markup Actually Does

What structured data is, and what it is not

Structured data is a block of code, usually in a format called JSON-LD, that describes what a page is about using a shared vocabulary from Schema.org. Instead of leaving a search engine to infer that the street name in your footer is your address and “Mon to Fri, 8 to 5” are your hours, the markup states it explicitly: this is a LocalBusiness, this is its address, these are its opening hours, this is its logo, these are its official profiles elsewhere.

What it does well:

Removes ambiguity about facts that are easy to misread: names, addresses, hours, dates, prices, authors.
Connects your website to your other official profiles, so systems understand they describe the same organization.
Makes pages eligible for certain enhanced search features, such as breadcrumbs, event listings, product information and article details, where the search engine supports them.

What it does not do:

It does not make a weak page strong. Markup describes content; it does not replace it.
It does not guarantee any enhanced display. Search engines decide when to use it.
It is not a ranking shortcut, and it is not a way into AI answers. Google has stated that its AI Overviews and AI Mode require no special schema or optimization beyond normal search guidance. Other AI assistants mostly retrieve pages and read the visible text; their providers do not document a role for markup in choosing what to cite.

So the honest position is the one on our SEO and AI visibility page: we implement structured data because it helps machines understand a business accurately, not because it is magic. The broader question of which facts machines need, and where they should live, is covered in what business information helps AI search systems.

Why it still matters in an AI search world

If markup does not earn citations directly, why bother? Because AI answers are built on search, and search engines use structured data to build their understanding of entities: which organization is which, where it is, what it does, how its pages relate. That understanding feeds results, business panels and knowledge used by AI features on top of them.

Think of it as reducing the chance of error rather than increasing the chance of selection. A business with two similar names in the region, a recent rebrand, a French and an English trading name, or several locations is exactly the kind that gets confused by machines. Clear Organization and LocalBusiness markup, with consistent identifiers and links to official profiles, makes those confusions less likely. When an assistant does mention you, it is more likely to get your details right.

How assistants find and weigh sources more generally is explained in how AI search engines find and recommend businesses, and the platform differences in how to show up in every AI search engine.

The types that matter for most businesses

Schema.org has hundreds of types. Most business websites need a handful, done well.

Organization

On the home page or about page: official name, alternate names (including a French name if you use one), logo, URL, contact points and sameAs links to your official profiles (LinkedIn, Business Profile, major directories, Wikipedia or Wikidata if they exist). This is the anchor that ties everything else together.

LocalBusiness, or a more specific subtype

For each physical location: a specific type where one fits (Dentist, LegalService, Restaurant, HomeAndConstructionBusiness and many others), with address, geo-coordinates, phone, opening hours including seasonal changes, and a link back to the parent organization. One location per page, with markup that matches that page.

Article or BlogPosting

For editorial content: headline, publication and modification dates, author (a person or the organization) and publisher. It helps search engines understand freshness and authorship, which matters for content intended to be cited.

BreadcrumbList

Describes where a page sits in the site’s structure. Simple, widely supported and useful for search results.

Where relevant: Product, Event, Service

Product markup for anything sold online, Event markup for dated events open to the public, and Service markup to describe what a service business offers. Service markup does not produce a special search display, but it is a clean way to describe offerings and the areas served.

Types to handle carefully

Some markup is widely recommended in older guides and no longer does what those guides claim.

FAQPage. Google limited FAQ rich results to well-known, authoritative government and health websites. For most businesses the markup no longer produces a visible result. It is harmless if it accurately reflects a visible FAQ, but adding it to every page in the hope of AI citations is wasted effort.
HowTo. Google no longer shows HowTo rich results. Not worth adding for search purposes.
Review and AggregateRating about yourself. Google does not show review stars for a LocalBusiness or Organization based on reviews it collects and displays about itself on its own site. Marking up your own testimonials will not produce stars, and inflating ratings in markup can lead to a manual action.
Speakable. A specialized, limited feature aimed at news publishers, not a general voice search tool for businesses.

How to implement it properly

Start from a canonical record. Write down the agreed facts about your organization and each location before writing any markup, so the code, the page and your listings all say the same thing.
Use JSON-LD. It is the format Google recommends, it sits separately from the visible HTML, and it is easier to maintain than inline formats.
Match the visible page. Markup must describe content a visitor can see on that page. Hidden facts, different hours or services that appear only in code break Google’s guidelines and undermine trust.
Generate it from your content, not by hand, where possible. On a custom site, build markup from the same data that renders the page, so a change to hours updates both. On WordPress and similar platforms, SEO plugins handle the basics; check what they produce rather than assuming.
Connect the entities. Give the organization a stable identifier and reference it from each location and article, so the pieces describe one coherent business rather than a pile of unrelated fragments.
Validate. Use Google’s Rich Results Test for features Google supports and the Schema Markup Validator for general correctness, then watch the enhancement reports in Search Console for errors over time.

Structured data belongs in a site build, not bolted on afterwards. It is part of every website we build through our website design and development work, and it is one of the items in the technical SEO audit checklist for existing sites.

Common mistakes that cause problems

Stale data. Hours, addresses or phone numbers in the markup that no longer match the page or the Business Profile. This happens constantly after moves and rebrands.
Every location on every page. One block listing all locations on every page, instead of each location’s markup on its own page.
Duplicate, conflicting blocks. A theme, a plugin and a developer each adding their own Organization markup with different details.
Wrong type. A law firm marked up as a generic Organization, or a restaurant marked up as a LocalBusiness with no subtype, when a precise type exists.
Markup for content that is not there. FAQ markup with no visible FAQ, or ratings that appear nowhere on the page.
Missing French equivalents. On bilingual sites, French pages with English markup, or with no markup at all. Each language version should describe itself in its own language.

Measuring what it does

Be realistic about measurement. You can see whether markup is valid and detected (Search Console enhancement reports), whether pages are eligible for and appearing with enhanced features, and whether your business information is displayed correctly in search. You cannot isolate the effect of markup on AI citations, because no assistant reports why it chose a source, and markup is rarely the only thing that changed.

The Squeeze, a restaurant client, is a fair example of where markup sits in practice: it was one part of a complete foundation (site, analytics, pixels, schema and content clusters) that took the restaurant from little visibility to steady organic growth. The case study has the figures. No one would credit the schema alone, and that is the right way to think about it: structured data is the accurate label on a good product, and the product is still what people buy.