FAQ schema in 2026: signal or noise

FAQ schema in 2026 is structured data that pairs a question with an answer in a machine-readable block, and it now serves exactly one purpose: helping an engine understand and extract your content, not winning you a visible feature in Google Search. The SERP reward is gone. The extraction value is not. Knowing which of those you are chasing is the whole decision.

The reason to revisit this now is that the deprecation finally finished. Google had restricted FAQ rich results to government and health sites back in 2023, but the markup still produced the familiar expandable Q-and-A in many results. That ended this year. Google's documentation added a notice that the feature "will no longer appear in Google Search starting May 7, 2026," and then the documentation page itself was removed in June with the note that the feature "is no longer shown in Google Search results" (Google Search Central).

The cleanup runs on a schedule. FAQ rich results stopped appearing on May 7, 2026. In June 2026, Search Console loses the FAQ search-appearance filter, the rich-result report, and FAQ support in the Rich Results Test. In August 2026, the Search Console API stops returning FAQ rich-result data (Search Engine Journal). Google did not publish a blog post or a reason, which is its usual pattern for retiring a search-appearance feature.

#What did not change

Two things survived the deprecation, and they are the reason this is not a "rip it all out" moment.

First, the markup is still valid. FAQPage remains a real Schema.org type, and Google's guidance is that existing markup "can stay in place. The markup won't cause problems, but it also won't produce visible results in Google Search" (Search Engine Journal). You are not being penalized for keeping it. You are simply no longer paid for it in the SERP.

Second, AI answer engines read content the way FAQ structure is built. An LLM doing retrieval against a user's question wants three things: the question stated explicitly, the answer short and self-contained, and the pairing unambiguous. A {question, answer} block delivers all three. That is why FAQ-structured content tends to be cheap for an engine to match and cheap to quote. None of that depended on Google's rich result, so none of it went away on May 7.

So the value moved. It used to be split between a visible Google feature and machine understanding. Now it is entirely machine understanding, and the audience for that understanding is increasingly ChatGPT, Perplexity, Gemini, and Google's own AI Overviews rather than the blue-link SERP.

#Where it is still signal

FAQ schema earns its place when the content under it is genuinely a question and answer that a real person asks.

• Help center and support articles whose primary content is "how do I do X" and a direct answer. This is the strongest case. The schema just makes explicit what the page already is.
• A short FAQ block on a feature or pricing page that surfaces real buyer questions: "Does this integrate with Salesforce?" or "Can I change plans later?" Mark only the genuine ones.
• Documentation built from real support tickets, where the question is phrased the way customers phrase it ("can I export my data?") rather than the way a marketer would ("what data-export functionality is available?").

The pattern in all three: the markup describes content that would deserve to be cited even if the schema were absent. The schema makes it easier to find and lift, it does not manufacture value that is not there.

#Where it is noise

The noise cases are the ones worth cutting, because the only thing they ever bought was the SERP feature that no longer exists.

The classic example is the marketing FAQ: three softball "questions" bolted onto every product page, with answers that are sales copy. "Why is your product the best?" is not a question an answer engine will ever serve, and a page stuffed with that pattern reads as low-signal. Google learned to discount it, and AI engines, trained on the same kind of judgment, treat it the same way.

Other noise to retire:

• FAQ blocks generated by SEO tooling that do not reflect questions any customer actually asked.
• Boilerplate "Why choose us?" framings duplicated across dozens of pages.
• FAQ schema on a blog post that is not actually a Q-and-A. A post should use Article (or BlogPosting) schema; add FAQPage only if the post genuinely contains a question-and-answer section.

If you added FAQ markup purely to win the expandable Google result, that investment has stopped paying out. Keeping it costs nothing, but maintaining it, or worse, generating more of it, is effort spent on a feature that is gone.

#The decision

Run one test on every FAQ block you own: if a person searched this exact question, would the answer below it deserve to be the response? If yes, keep the schema and keep it accurate, because that is precisely the content an answer engine wants to extract. If no, the block was always about gaming a SERP feature, and that feature is now retired, so the block is noise.

This is the same instinct Google has applied for years to structured data generally: markup is meant to describe content that exists, not to conjure a ranking signal out of thin air. The deprecation just made the dividing line obvious. The teams that wrote real FAQ content lost nothing on May 7. The teams that wrote FAQ markup as a SERP trick lost the only thing they were getting.

FAQ schema is not dead, and it is not a growth lever either. It is a small, free clarity signal that pays off when, and only when, the question and answer underneath it are real.