Schema markup for AI search: what still matters in 2026
Hephaestus Technology6 min read

Schema markup is not required for AI search: Google states that there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode. Structured data still matters in 2026 for rich results in classic search and for describing your business accurately, so the practical job is to keep a small set of correct, maintained markup rather than chase "AI schema".
Key takeaways
- Google states that structured data is not required for its AI features and that there is no special schema.org markup to add for them.
- Structured data still matters for rich results in classic search, and Google recommends JSON-LD as the easiest format to maintain.
- Several rich result types have been cut back: FAQ results are limited to authoritative government and health sites, HowTo results are gone, and seven more types were phased out in 2025.
- Markup must match the visible content of the page; inaccurate schema is a liability, not a shortcut.
- OpenAI, Anthropic and Perplexity do not document schema as a factor in their crawler guidance, so treat claims about "AI schema" with caution.
What is schema markup?
Schema markup, or structured data, is code added to a page that labels its content in a standard vocabulary from schema.org. Google defines it as "a standardized format for providing information about a page and classifying the page content" in its introduction to structured data. It recommends JSON-LD in most cases because it is easiest to implement and maintain, while saying Microdata and RDFa are equally fine if valid.
Structured data can make pages eligible for rich results, such as product details or article information, but it does not guarantee them.
Does schema markup help with AI search?
Not in the way many vendors suggest. Google's AI features documentation says "there's also no special schema.org structured data that you need to add" for AI Overviews and AI Mode. Its optimisation guide for generative AI goes further, listing "overfocusing on structured data" among things you can ignore, while still recommending structured data as part of your overall SEO strategy because it supports rich results.
For other assistants, there is no published guidance either way. The crawler documentation from OpenAI, Anthropic and Perplexity explains how to allow or block their bots but says nothing about schema affecting citations. Anyone claiming a proven schema effect in ChatGPT or Claude answers should be able to show you the evidence.
The honest summary: schema is part of good SEO hygiene, and SEO is the foundation of AI visibility, as we explain in GEO vs SEO for B2B websites. It is not a lever on its own.
Which structured data features have been cut back?
Google has reduced several rich result types, so some older advice is out of date.
| Change | Date | What it means | Source |
|---|---|---|---|
| FAQ rich results limited to well-known, authoritative government and health sites | August 2023 | Most B2B sites no longer get FAQ snippets | Google Search Central |
| HowTo rich results limited to desktop, then removed | August to September 2023 | HowTo markup has no visible effect in Search | Google Search Central |
| Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement and Vehicle Listing phased out | June 2025 | Fewer visual enhancements; "won't affect how pages are ranked" | Google Search Central |
Google said in 2023 that there is "no need to proactively remove" unused markup, which "does not cause problems for Search, but also has no visible effects".
Step 1: How do you audit the schema you already have?
Start by finding out what markup your site already outputs, because themes and plugins often add types nobody chose.
- Run your home page, a service page and a blog post through Google's Rich Results Test and the Schema Markup Validator.
- Check the enhancement reports in Search Console for errors and warnings.
- List every type found, which template outputs it, and whether its content matches the visible page.
Most small B2B sites find a mix of useful Organization markup, plugin-generated Article markup and leftover FAQ or HowTo markup from older SEO advice.
Step 2: What should a B2B site keep or add?
Keep a small, accurate set tied to real content:
- Organization: name, logo, URL, contact details and social profiles on the home or about page. This describes your business consistently wherever it is parsed.
- Article or BlogPosting: author, publication and modification dates on articles. Dates and authorship help readers judge freshness, whether shown as markup or on the page.
- Product or SoftwareApplication: only if you sell a product with visible price or review information that matches the markup.
- BreadcrumbList: helps show site structure in results.
- FAQPage: optional. Harmless if accurate, but do not expect rich results unless you are a government or health site.
Google reports past case studies on its structured data page, such as Rotten Tomatoes measuring a 25% higher click-through rate on pages enhanced with structured data. These are rich result gains in classic search, published by Google, and not evidence of AI citation effects.
Step 3: How do you keep markup accurate?
Generate markup from the same data that renders the page, so prices, dates and names cannot drift apart. Google's May 2025 guidance on AI search repeats the rule: make sure that all the content in your markup is also visible on the page, and validate it. Markup that contradicts the visible page is worse than none, because it can make pages ineligible for rich results.
Step 4: What should you not spend money on?
Skip anything sold as special "AI schema", llms.txt-style markup bundles or bulk FAQ generation. Google says you do not need new machine-readable files or markup to appear in its AI features, and our explainer on what llms.txt is covers the file most often sold alongside schema packages. The money is better spent on the page content itself; our content for AI citations checklist shows where.
Your schema checklist for 2026
- [ ] Rich Results Test run on key templates
- [ ] Search Console enhancement reports checked for errors
- [ ] Organization markup present and accurate
- [ ] Article markup includes real author and dates
- [ ] Product or pricing markup matches visible prices
- [ ] Leftover HowTo and FAQ markup reviewed (keep if accurate, do not expand)
- [ ] Markup generated from the same data as page content
- [ ] No "AI schema" packages or hidden content
Frequently asked questions
Does schema markup help my site appear in AI Overviews?
Not directly, according to Google. Its guidance says structured data is not required for generative AI search and that there is no special schema to add. It still recommends structured data as part of an overall SEO strategy because it supports rich results.
Should I remove FAQPage schema now that FAQ rich results are restricted?
There is no need. Google said in 2023 that there is no need to proactively remove it, and that unused structured data does not cause problems for Search. Keep it if it is accurate and easy to maintain; do not add it expecting rich results.
Does ChatGPT read schema markup?
OpenAI does not say so in its crawler documentation, and neither does Anthropic or Perplexity. Some AI tools may parse structured data when they fetch a page, but there is no published guidance that it affects citations. Focus on clear visible content first.
Which schema type should a small B2B company start with?
Organization markup on the home or about page is a sensible baseline, with accurate name, logo, URL and contact details. Then add types that match real page content and are eligible for rich results in Google, such as Article or Product where relevant.
Can incorrect schema hurt my site?
Yes. Google requires that markup reflects content visible on the page, and markup that does not follow its guidelines can make a page ineligible for rich results.
If you want help auditing your site's markup and content together, book a free 30-minute call with our team.
Cover photo by Lukas Blazek on Pexels.