SEO markup for AI: how to structure content for search engines

SEO markup is schema.org structured data, usually in JSON-LD, that tells search engines and AI assistants who wrote the page, which organisation stands behind it and what product or service it describes. Markup clarifies facts. Rankings still come from content.

What is SEO markup and why does it matter in 2026?

Structured data describes a page in the schema.org vocabulary. A visitor reads "Marketing MIX, Kyiv, phone number". A crawler gets the same fact explicitly: type Organization, address, telephone, links to official profiles.

Google uses markup for rich results: product price and availability, ratings, breadcrumbs, the organisation panel. AI systems (AI Overviews and AI Mode in Google, ChatGPT Search, Perplexity, Copilot) read the same HTML. Clear entities make it easier for them to connect a brand, an author, an address and a service.

Google's documentation on AI features in Search is explicit: there is no special "AI markup", and normal SEO requirements apply. So markup is technical hygiene. It will not carry a site on its own.

Which schema.org types does a website need?

Every extra type is another place for errors. Start with this set and add more only where it earns its place.

Why is JSON-LD easier than Microdata?

JSON-LD sits in its own script type="application/ld+json" block, separate from the layout. It is easier to generate from a CMS, test and update. Microdata lives inside HTML attributes and breaks with every redesign.

One rule matters most: one entity, one description. If a WordPress theme already outputs Organization, a plugin must not add a second one with a different phone number. On MODX and other CMSs we build markup with a snippet from page fields, so it never drifts from the visible text.

How do you connect entities with @id?

Strong markup is a graph. The organisation gets a permanent @id, for example the home URL with #organization. An article references it as publisher, a service as provider. Search engines then connect every mention to one company.

For authors, use Person with a job title and profile links. That supports the experience signals (E-E-A-T) described in Google's quality rater guidelines.

How do you test structured data?

1. Rich Results Test by Google: are rich results eligible, are there critical errors.
2. Schema Markup Validator (validator.schema.org): syntax and vocabulary.
3. Search Console, Enhancements reports: site-wide errors after indexing.
4. Manual check: price, address and opening hours in JSON-LD match the visible page.

What are the most common markup mistakes?

• Review and rating markup for reviews that are not on the page.
• The same FAQPage copied across dozens of pages.
• Two Organization blocks with different names or phone numbers.
• Retired types: Google dropped HowTo rich results in 2023 and several rare rich result types in 2025.
• Markup on pages blocked from indexing, or blocked for AI crawlers in robots.txt (OAI-SearchBot, PerplexityBot, Bingbot).

What else helps AI search understand a site?

Markup works together with the copy. A direct answer in the first paragraph, question headings, fact tables, a visible update date and a clear author give an AI assistant a ready passage to cite. llms.txt is a proposed table-of-contents file for language models. Google does not officially use it, and it does no harm.

This is the core of our website promotion in AI search results service: markup audit, entity graph, question-and-answer copy and tracking brand mentions in ChatGPT and Gemini answers. The classic side is covered by search engine promotion (SEO). In our experience, first noticeable results arrive in about 3 months.

FAQ

Does schema.org markup improve Google rankings?

It is not a direct ranking factor. It enables rich results that can raise click-through rate, and it helps search engines understand the entities on a page.

Do ChatGPT and AI Overviews need special markup?

There is no separate standard. AI systems read the same HTML and JSON-LD. It matters more that the page is open to their crawlers and contains direct answers.

Which structured data format does Google recommend?

Google recommends JSON-LD. Microdata and RDFa are supported too, but they are harder to maintain when the layout changes.

Should I add FAQPage if Google no longer shows FAQ snippets?

Yes, when the page has a real Q&A block. It helps AI answers. Copying the same FAQ across many pages does not.

How often should markup be checked?

After every redesign, theme or plugin update, and quarterly through Search Console reports.

We will audit your structured data and build an entity graph for Google and AI search. Send a request and we will reply during business hours.

Request a markup auditTelegramWhatsApp