A founder spends an afternoon adding FAQPage JSON-LD to every service page, following a guide that promised it was the fastest way to get cited by ChatGPT and Google AI Overviews. Weeks later, no rich snippet shows up in search, no citation shows up in ChatGPT, and the conclusion is that schema markup doesn't work anymore. The premise was already off: the rich snippet in Google Search and a citation inside an AI-generated answer are two different systems, and in May 2026, Google removed one of them from search entirely, which is why the confusion is spreading right now.
What schema markup actually does, and what it doesn't
Schema markup, usually written in JSON-LD, is a machine-readable layer that sits alongside your visible page content and explicitly labels what things are, this is an Organization, this is a Person, this is an Article with this publish date. It doesn't rank content higher and it doesn't inject anything into a generated answer, it disambiguates what's already on the page for a system that can't infer entity relationships from prose the way a person can.
That distinction matters because most of the advice circulating treats schema as a ranking or citation lever you can pull, when it's closer to labeling a diagram so a machine doesn't have to guess what each part is.
The May 2026 FAQ rich-results deprecation, and why it doesn't kill FAQ schema
Google added a deprecation notice to its FAQ structured data documentation and FAQ rich results stopped appearing in Google Search results starting May 7, 2026, followed by the removal of the FAQ search-appearance filter and Rich Results Test support in June, according to Search Engine Journal's coverage of the change. If you were adding FAQPage schema specifically to win that visual snippet in search results, that specific payoff is gone.
What didn't change: FAQPage remains a valid schema.org type, and having it on your pages doesn't cause any problem, it simply no longer triggers the rich-result display it used to. Ripping it out isn't necessary, and it's not the actual question worth answering next, which is whether schema was ever the thing driving AI citations in the first place.
What Google actually confirmed about structured data and AI Overviews
There is no special schema.org markup required for AI Overviews or AI Mode eligibility, per Google Search Central's documentation on how structured data works, the same core content and SEO practices apply. Structured data can help Google understand a page more precisely, which can support better retrieval and understanding, but it is not confirmed as a direct citation trigger, and no schema type guarantees inclusion in an AI-generated answer.
The honest takeaway is less exciting than most schema-for-AI guides suggest: structured data is a clarity layer that helps a system correctly parse what's already on your page, not a hack that manufactures a citation out of thin content. If the underlying prose doesn't answer the question directly, schema around it won't make it get quoted.
Where schema still genuinely earns its keep for AI visibility
- Organization and Person schema resolve who is actually speaking, tying your brand and author identity together consistently across your site, which matters for the kind of entity understanding AI systems increasingly rely on for trust signals.
- Article and BlogPosting schema carry datePublished and dateModified fields, one of the few structural freshness signals a crawler can read without inferring it from prose, which connects directly to how often you need to update content to stay cited.
- LocalBusiness schema disambiguates location and service entities for local AI answer surfaces, useful for any business competing on 'near me' or city-specific queries.
- None of this replaces crawler access or answer-first prose, schema is the machine-readable half of a system that also needs the llms.txt and crawler access layer and the answer-first paragraph structure working alongside it, not instead of it.
How AIBOOTSTRAPPER solved this for a client
PropLock, a UK real estate platform we built, paired correct entity and article schema with AI-crawler access and answer-first listing copy from day one, not as separate afterthought tasks. The site reached 12,000 organic visitors a month within 90 days of launch and a 47% increase in qualified viewings, a result of the full system working together, not any single tag.
How AIBOOTSTRAPPER helps
AIBOOTSTRAPPER builds GEO optimized websites with the full stack in place from the first line of code, correct entity schema, crawler access, and answer-first content structure, rather than treating schema as a checkbox that's supposed to work on its own.
If you've added schema markup and still aren't seeing AI citations, book a call and we'll audit whether the gap is structural, content-level, or both.
Want this done for you?
Book a free strategy call and we'll show you how to build and market your business with AI.