Schema for AI search has become unnecessarily complicated.
Schema.org currently contains more than 800 types. That does not mean a typical UK SME needs hundreds of them. Schema.org’s current development vocabulary lists 826 types and more than 1,500 properties. Schema.org
The other mistake is assuming that adding more structured data automatically makes ChatGPT, Gemini or Google’s AI features more likely to recommend your business.
There is no published evidence supporting that.
Google is particularly clear. Its 2026 generative AI guidance says structured data is not required for generative AI search and there is no special Schema.org markup you need to add. Existing SEO practices still apply, and structured data remains useful as part of the wider search ecosystem. Google for Developers
So this is not a schema encyclopaedia.
It is a prioritisation guide.
For a typical service-based SME, there are around six types I would inspect before worrying about anything more exotic:
- Organization
- LocalBusiness
- Service
- Article
- BreadcrumbList
- FAQPage, with an important 2026 caveat
The right combination depends on what the business and page actually are.
Table of Contents
- What schema actually does
- The six schema types worth checking first
- Organization schema
- LocalBusiness schema
- Service schema
- Article schema
- Breadcrumb schema
- FAQPage schema and the 2026 problem
- Which schema types should SMEs stop obsessing over?
- JSON-LD or microdata?
- How to validate schema properly
- Does AIOSEO handle schema?
- The honest limit of schema for AI search
- Frequently Asked Questions About Schema and AI Search
What schema actually does
Structured data gives machines explicit information about what a page represents.
Google describes structured data as a standardised format for providing information about a page and classifying its content. Google for Developers
For example, markup can help describe:
This is the organisation.
This is its address.
This page describes a service.
This person wrote this article.
This page sits underneath this section of the website.
That can remove ambiguity.
But schema does not replace the page itself.
If your service page contains vague copy, adding:
“@type”: “Service”
does not suddenly make the proposition clear.
If your business information is inconsistent across the web, Organization markup does not erase those contradictions.
If the content is weak, Article schema does not make it authoritative.
Think of structured data as:
classification and clarification
rather than:
a shortcut to recommendation visibility.
Google itself says structured data provides explicit clues about page meaning, while also warning that even valid structured data does not guarantee enhanced search appearances. Google for Developers
Google’s introduction to structured data
The six schema types worth checking first
For a typical service company, this is the practical shortlist.
| Schema Type | What It Helps Describe | Typical Priority |
|---|---|---|
| Organization | Who the company is | High |
| LocalBusiness | A physical or local business and its operating details | High where relevant |
| Service | What the company provides | High for service companies |
| Article | A blog, article or editorial piece | High for content |
| BreadcrumbList | Where the page sits in the site hierarchy | High |
| FAQPage | A page containing multiple FAQs | Conditional |
This is not an AI ranking-factor table.
It is a practical implementation order.
Different sites need different types.
An ecommerce shop should care about Product markup.
A recruiter may need JobPosting.
An events company may need Event.
Do not ignore the schema that directly describes your business model simply because it is not on this list.
1. Organization schema
For most businesses, start here.
Organization structured data describes the company itself.
It can include information such as:
- organisation name
- URL
- logo
- contact details
- address
- identifiers
- social or external profiles
Google says Organization markup can help it understand administrative details about an organisation and disambiguate that organisation in Search. Google for Developers
That makes the purpose straightforward.
It helps answer:
Who is this business?
What to check
Make sure your schema agrees with the visible website.
Check:
Business name
Canonical URL
Logo
Contact information
Address where appropriate
Relevant external profiles
Do not use schema to introduce information that contradicts what users see.
Google specifically advises that structured data should represent the visible page content. Google for Developers
Google’s Organization structured data guidance
2. LocalBusiness schema
If you serve customers from a physical location or operate a genuinely local business, LocalBusiness deserves attention.
Schema.org defines LocalBusiness as both an Organisation and a Place, which means it inherits relevant properties from both. Schema Markup Validator
Google supports LocalBusiness structured data for details including:
- business address
- opening hours
- departments
- telephone
- geographic information
It can also contribute to Google’s understanding of businesses appearing in Search and Maps. Google for Developers
It helps answer:
Where is this business and how does it operate locally?
Important warning
Do not create fake locations to manufacture local relevance.
If you have one real office and serve twenty cities, that does not mean you have twenty LocalBusiness entities.
Your schema should describe reality.
Google’s LocalBusiness structured data guidance
3. Service schema
This is particularly useful conceptually for B2B and service companies.
Schema.org defines Service as a service provided by an organisation. Its available properties include things such as:
- serviceType
- provider
- areaServed
- hasOfferCatalog
- hoursAvailable
- hasCertification Schema.org
It helps describe:
What does this company actually provide?
Imagine a facilities management company.
The Organisation tells us:
ABC Facilities Ltd
The Service entity can then describe:
Planned and reactive maintenance
and connect it back to the provider.
For a local service, it can also communicate the area served.
That is useful semantic structure.
But there is an important distinction
Google does not currently list generic Service as one of its dedicated rich-result experiences in the way it supports Article, Product or Breadcrumb markup. Google’s Search Gallery documents the structured-data features it directly supports for Search appearances. Google for Developers
So do not sell Service schema to a client as:
This will create a special Google result.
Its value is primarily descriptive.
Schema.org Service specification
4. Article schema
If you publish:
- blogs
- guides
- research
- analysis
- news
- thought leadership
Article markup deserves to be configured properly.
Google supports:
- Article
- NewsArticle
- BlogPosting
and recommends information including the author and author URL to help it understand authorship. Google for Developers
This helps answer:
What is this content, and who published or wrote it?
For a B2B website, useful fields can include:
- headline
- author
- author URL
- date published
- date modified
- image
Do not automatically mark every page as Article.
A commercial service page is not a blog post just because it contains 1,500 words.
Match the schema to the purpose of the page.
Google’s Article structured data guidance
5. BreadcrumbList schema
Breadcrumb markup is less glamorous.
That is precisely why it often gets ignored.
Breadcrumbs communicate where a page sits within the website hierarchy.
For example:
Home
→ Services
→ GEO & AEO
→ AI Visibility Audit
Google says breadcrumb trails indicate the page’s position in the site hierarchy and help users understand and navigate that hierarchy. Google for Developers
That helps answer:
Where does this page belong?
For larger B2B websites with:
- sectors
- services
- locations
- resources
- case studies
clear hierarchy matters.
Breadcrumb schema supports that structure.
Google’s Breadcrumb structured data guidance
6. FAQPage schema comes with a major 2026 caveat
This is where old schema advice can become misleading.
FAQ markup used to be discussed heavily because Google could show FAQ rich results directly in organic listings.
That is no longer the reason to prioritise it.
Google deprecated FAQ rich results in May 2026 and subsequently removed its FAQ rich-result documentation in June. Google for Developers
So if somebody tells you:
Add FAQ schema so Google gives you a giant FAQ result.
that advice is outdated.
Should you remove every FAQPage schema implementation?
Not necessarily.
FAQPage still exists within the wider Schema.org vocabulary, and CMS tools can still generate it.
But for the average SME, it is no longer a high-priority Google rich-result tactic.
Treat it as conditional.
Your visible FAQ content itself can still be extremely useful because it answers customer questions clearly.
The mistake is confusing:
useful FAQ content
with:
FAQ schema guarantees additional search visibility.
It does not.
This is also why I would rank FAQPage below the other five for most SMEs in 2026.
Which schema types should SMEs stop obsessing over?
There is no single schema type that is universally a waste of time.
The waste comes from adding markup that has no meaningful relationship to the business or page.
Common examples include:
Marking everything up because a plugin offers it
More schema is not automatically better schema.
Using the wrong page type
A service page is not an Article just because WordPress calls it a post type.
Adding FAQPage solely for an old Google rich-result tactic
That specific Google feature is gone. Google for Developers
Creating review markup that does not reflect legitimate visible review content
Google’s structured-data policies require markup to accurately represent the page and prohibit misleading implementations. Google for Developers
Adding exotic types because they look sophisticated
If nobody consuming your structured data benefits from the additional distinction, you have probably found a lower-priority task.
Duplicating schema from multiple plugins
This happens frequently on WordPress.
One plugin outputs Organization.
Another outputs LocalBusiness.
The theme adds another graph.
Custom code adds a fourth.
The result can become harder to maintain rather than clearer.
JSON-LD or microdata?
For most businesses:
Use JSON-LD.
Google supports:
- JSON-LD
- Microdata
- RDFa
but its structured-data guidelines identify JSON-LD as the recommended format. Google for Developers
JSON-LD is usually easier to:
- generate
- read
- edit
- debug
- maintain
because it sits in a structured script block rather than being spread across individual HTML attributes.
A simple example looks like this:
<script type=”application/ld+json”>
{
“@context”: “https://schema.org”,
“@type”: “Organization”,
“name”: “Example Ltd”,
“url”: “https://example.co.uk/”
}
</script>
That does not mean manually adding JSON-LD is always the right approach.
If WordPress or your SEO plugin already outputs correct markup, avoid creating unnecessary duplicate schema.
How to validate schema properly
Do not trust the green tick inside your WordPress plugin.
Test the output.
There are two useful tools, and they answer slightly different questions.
Google Rich Results Test
Use this when you want to know whether markup qualifies for a Google-supported rich-result feature.
Google recommends the Rich Results Test for supported Search features. Google for Developers
Schema.org Markup Validator
Use this when you want to validate broader Schema.org markup, including types that Google does not provide a dedicated rich result for.
Schema.org says its validator can extract and validate:
- JSON-LD
- RDFa
- Microdata
and identify syntax problems. Schema.org
This distinction matters for something like Service.
The Schema.org validator can tell you whether the markup itself is valid.
Google’s Rich Results Test may not show a special Service rich-result feature because Google does not offer one.
Does AIOSEO handle this?
Yes, AIOSEO can handle a substantial amount of common schema work.
Its current documentation says it outputs JSON-LD and supports markup including:
- Organization
- Local Business
- Article
- Breadcrumb
- FAQ
- Service
- Product
- Event
- Job Posting
among other types. All in One SEO
Its 2026 release also introduced an AI Schema Generator for Pro users. All in One SEO
That means many WordPress SMEs do not need a developer manually coding every schema graph.
But the plugin cannot decide your business reality for you.
You still need to check:
- correct organisation details
- correct schema type
- correct address
- current phone number
- valid social profiles
- whether the visible content supports the markup
- whether another plugin is already generating the same schema
A technically valid schema graph containing inaccurate business information is still inaccurate.
AIOSEO’s Schema.org markup guide
The honest limit of schema for AI search
This is the section worth remembering.
Schema can make information explicit.
It cannot manufacture the information.
Organization markup can describe your organisation.
It cannot give the organisation credibility it has not earned.
Service markup can describe your services.
It cannot make a thin service page useful.
Article markup can identify the author.
It cannot turn generic content into expertise.
LocalBusiness markup can clarify location details.
It cannot compensate for incorrect business profiles across the wider web.
FAQ markup can describe questions and answers.
It cannot guarantee that an AI assistant uses those answers.
And no schema type guarantees that ChatGPT recommends you.
Google’s own 2026 generative AI guidance explicitly warns against overfocusing on structured data. There is no special schema required for Google’s generative AI search features. Google for Developers
That makes the sensible order:
Get the content right.
Get the entity information right.
Get the site architecture right.
Then describe those things accurately with structured data.
Not the other way around.
For more on structuring information so it can be understood and reused:
Learn more about content structures that generative AI engines can understand
For the wider AEO approach:
Frequently Asked Questions About Schema and AI Search
Does schema guarantee an AI Overview or rich result?
No.
Google explicitly says valid structured data does not guarantee that a rich result will appear.
Google structured data policiesGoogle also says there is no special structured data required to appear in AI Overviews or AI Mode.
Google AI features guidanceSchema creates eligibility and helps describe information. It does not guarantee placement.
Does schema make ChatGPT recommend my business?
There is no published evidence supporting that claim.
OpenAI has not published a business-recommendation formula saying that adding Schema.org markup causes a company to be selected.
Use schema because it makes website information more explicit and supports established search practices.
Do not sell it as a ChatGPT recommendation switch.
Does AIOSEO handle schema automatically?
It handles many common schema types and outputs JSON-LD.
But automation does not remove the need to verify the resulting graph.
Check the live page and make sure the information matches the business and visible content.
AIOSEO Schema.org markup guideJSON-LD or microdata?
For most sites, use JSON-LD.
Google supports JSON-LD, Microdata and RDFa but recommends JSON-LD.
Google structured data guidelinesIt is generally easier to maintain and debug.
How do I validate schema?
Use two tools where appropriate:
Google Rich Results Test
For structured data that supports Google Search features.
Schema.org Markup Validator
For broader Schema.org validation.
Do not rely solely on the validation message inside your SEO plugin.
Should I still use FAQ schema in 2026?
You can, where it accurately describes genuine FAQ content.
But do not implement it expecting Google’s old FAQ rich results.
Google deprecated that search feature in May 2026 and removed the associated documentation in June.
Google Search documentation updatesFor most SMEs, creating genuinely useful FAQ content is more important than obsessing over the FAQPage markup itself.
Do I need all six schema types on every page?
No.
That would defeat the point.
Use the types that accurately describe the page.
Blog article:
Article + BreadcrumbList
Local service page:
Organization / LocalBusiness + Service + BreadcrumbList
Do not add schema just to increase the number of types detected.
Start with six. Add more only when the business needs them.
You do not need to master 826 schema types.
You need to describe your business and its important pages accurately.
For most service-based SMEs, start by checking:
Organization
LocalBusiness where appropriate
Service
Article
BreadcrumbList
FAQPage only where useful and accurate
Then look at the business model.
If you sell products, investigate Product.
If you recruit, JobPosting may matter.
If you run events, Event may matter.
Schema should follow the actual content and business.
Not a checklist downloaded from an SEO blog.
And remember the limit:
Schema helps machines classify information.
It does not create authority, quality or recommendation eligibility by itself.
If you want to see whether your technical setup, content and entity information are helping or limiting your wider AI visibility:
Get your free AI visibility audit
For answer-focused optimisation:
For content structure:
