Does schema markup help AI search? What B2B sites should add
Schema markup does not get a B2B page cited in AI answers on its own. Google says there is no special schema for its generative AI features, and an Ahrefs study of 1,885 pages that added JSON-LD (vendor research) found no major change in AI citations, while Microsoft has said schema helps its models understand content. Keep a small, accurate set, mainly Organization, Article, BreadcrumbList and Product or SoftwareApplication where true, because it removes ambiguity; put the rest of the effort into content and sources.
Key takeaways
- Google: structured data is not required for generative AI search and there is no special markup for it.
- Ahrefs tracked 1,885 pages that added JSON-LD and saw no major citation uplift in AI Overviews, AI Mode or ChatGPT within 30 days.
- Microsoft has said schema helps its models understand content, so it still has a role for Bing and Copilot.
- Use schema to state facts precisely, never to describe things that are not on the page.
What the evidence says
Schema is the most oversold tactic in GEO. Plenty of agencies sell "AI schema" packages; the primary sources are more measured. Here is what each source actually says, and what kind of source it is.
| Source | What it says | Type |
|---|---|---|
| Google optimization guide for generative AI features (July 2026) | Structured data is not required for generative AI search, and there is no special schema.org markup to add; keep using it as part of overall SEO | Official documentation |
| Google, AI features and your website | Make sure structured data matches the visible text on the page | Official documentation |
| Microsoft, Fabrice Canel at SMX Munich | Schema markup helps Microsoft's models understand content | Official statement reported by Search Engine Land |
| Ahrefs, 1,885 pages adding JSON-LD (May 2026) | AI Overviews citations fell 4.6% against matched controls; AI Mode +2.4% and ChatGPT +2.2%, both statistically indistinguishable from zero | Vendor study |
The Ahrefs authors list real limits: the pages were already heavily cited, schema types were pooled, the window was 30 days and other changes may have happened at the same time. They also note that 53% of AI-cited pages have schema but attribute that to those sites having stronger SEO overall. The fair reading is that adding schema is not a lever for citations in the short term, not that schema is useless.
Why we still add schema on B2B sites
Structured data does three useful jobs. It states facts in a form machines do not have to guess: your legal name, logo, headquarters, product names and publication dates. It disambiguates: Google's Organization documentation says the markup helps it tell your organization apart from others. And it still powers search features Google supports, such as breadcrumbs, organization details in knowledge panels, review snippets, video and profile pages.
For AI search, the first two jobs matter most. An assistant that is sure which company you are and what you sell is more likely to describe you correctly. That is entity work, which our guide to entity SEO for B2B brands covers in depth.
Microsoft's position adds a practical reason. If schema helps Bing's models understand content, as Microsoft has said, then accurate markup supports Copilot answers, and Copilot is built into the Microsoft 365 tools that enterprise buyers already work in.
One change in 2026 matters for anyone who bought FAQ schema for rich results: Google says the FAQ rich result is no longer shown in Google Search, after the feature stopped appearing from May 7, 2026. FAQ markup on a visible FAQ is still valid schema.org and harmless, but it no longer earns a special result in Google.
The B2B schema set we recommend
Small, accurate and generated from the same fields as the visible page.
| Type | Where | Priority | Note |
|---|---|---|---|
| Organization | Home or About page | High | Name, legal name, logo, sameAs profiles, address, identifiers |
| BreadcrumbList | All pages below the home page | High | Matches the visible breadcrumb and URL structure |
| Article or BlogPosting | Articles and guides | High | Author, datePublished and dateModified matching the visible dates |
| Product or SoftwareApplication | Product pages | Medium | Only real offers and prices that appear on the page |
| Service | Service pages | Medium | Provider, service type and area served, as stated on the page |
| ProfilePage and Person | Team and executive pages | Medium | For people you want recognized as experts |
| VideoObject | Pages with embedded video | Medium | Title, description, upload date, transcript on the page |
| FAQPage | Pages with a visible FAQ | Low | No Google rich result since 2026; keep only if the FAQ is visible |
Get an AI visibility read across ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews.
What not to do
Schema causes more problems when it is used to say things the page does not.
- Marking up content that is not visible. Google warns against it even when the information is accurate.
- Adding ratings or reviews markup that does not come from real, visible reviews.
- Injecting schema with a tag manager and never checking that crawlers receive it.
- Letting schema dates drift from visible dates. Google asks for consistent dates on the page and in markup.
- Buying "AI schema" or llms.txt packages as a GEO strategy. Google says neither is needed for its generative AI features.
- Using different company or product names in schema than on the page and profiles.
How to implement it in a week
On most B2B sites this is a template job, not a page-by-page one.
- List the page templates on the site: home, about, service, product, article, team, location.
- Map each template to the schema types in the table above and to the CMS fields that already hold the values.
- Generate JSON-LD server-side from those fields, so markup and visible text cannot disagree.
- Validate a sample of each template with the Rich Results Test and the Schema Markup Validator.
- Check in Search Console that enhancement reports pick up the markup without errors.
- Add a QA check to your build or release process that fails when required fields are empty.
We run this as part of technical SEO. Our B2B SEO work covers templates, rendering and indexing together, because schema on a page Google cannot index does nothing.
How to measure the effect honestly
Do not expect a citation jump the week schema ships. Measure what schema can change: fewer errors in Search Console enhancement reports, correct logo and organization details in branded results, and more accurate descriptions of your company in a fixed monthly prompt set. If those improve and citations do not, the gap is content or sources, not markup.
When someone promises that schema alone will get you into ChatGPT or AI Mode, ask for the controlled test behind the claim. The best public one we know of found no major effect.
Where to start
Start with Organization, BreadcrumbList and Article, generated from templates, then add Product, Service and Profile page markup where the pages support them.
Our generative engine optimization service treats schema as one part of the entity and technical layer, alongside content and third-party sources.
If you want a second opinion on a schema package you have been offered, talk to our team.
Frequently asked questions
Does schema markup help you appear in ChatGPT?
There is no evidence that adding schema alone increases ChatGPT citations. In an Ahrefs study of 1,885 pages (vendor research), ChatGPT citations rose 2.2% after JSON-LD was added, a change statistically indistinguishable from zero.
Does Google use structured data for AI Overviews and AI Mode?
Google says structured data is not required for its generative AI features and there is no special schema to add. It recommends keeping structured data that matches visible content as part of overall SEO.
Does Bing or Copilot use schema?
Microsoft's Fabrice Canel confirmed at SMX Munich that schema markup helps Microsoft's models understand content. That makes accurate schema worth having for Bing and Copilot, though it does not guarantee citations.
Is FAQ schema still worth adding?
Google no longer shows FAQ rich results in Search. FAQ markup on a visible FAQ is still valid and harmless, but it is a low priority compared with Organization, Article and Product markup.
Which schema types matter most for B2B?
Organization, BreadcrumbList and Article first, then Product or SoftwareApplication, Service, ProfilePage and VideoObject where the pages support them.
Sources
- Google Search Central: Google's guide to optimizing for generative AI features on Google Search (updated July 10, 2026)
- Google Search Central: AI features and your website (updated December 10, 2025)
- Google Search Central: Introduction to structured data markup (updated December 10, 2025)
- Google Search Central: Organization structured data (updated September 8, 2026)
- Google Search Central: FAQPage structured data and the end of FAQ rich results (2026)
- Google Search Central: Structured data markup that Google Search supports (updated October 16, 2026)
- Google Search Central: Influencing your byline dates in Google Search
- Ahrefs (vendor study): We tracked 1,885 pages adding schema. AI citations barely moved (May 11, 2026)
- Search Engine Land: Microsoft Bing and Copilot use schema for their LLMs (Fabrice Canel, SMX Munich)
Related service: Generative engine optimization. See also pricing and client results.
We run enterprise B2B demand generation for teams in Montreal, Toronto, Ottawa and the United States.
