FAQ Schema Is Deprecated. What to Do With the Markup

Google stopped showing FAQ rich results on 7 May 2026. Five weeks later it deleted the documentation. Nothing about your markup broke, nothing needs removing, and nothing about your rankings changed. What you lost was the accordion of questions under your result, and the click-through those extra lines used to pull.

The honest answer to "what now" is more layered than either side of the usual argument. The people saying "rip it out" are solving a problem that does not exist. The people saying "keep it, AI reads it" are asserting something nobody has published evidence for. Here is what is documented, what is verifiable, and where the evidence stops.

The short answer

• The FAQ rich result stopped appearing on 7 May 2026. Google logged it in the documentation changelog on 8 May.
• The FAQ doc was removed on 15 June 2026. The old URL now redirects to the changelog entry announcing its own deletion.
• FAQPage is still valid schema.org with no deprecation notice. The markup did not become wrong.
• Google's position on retired markup, unchanged since 2023: unused structured data causes no problems for Search and has no visible effect.
• Nobody has published a measurement showing FAQPage markup changes AI citation rates. We have not measured it either.

What Google actually did, with dates

The primary source is Google's own documentation changelog, and it is short enough to quote in full.

The 8 May 2026 entry, headed "Deprecating the FAQ rich result feature":

What: Added a deprecation notice to the FAQ rich result documentation. Why: This feature will no longer appear in Google Search starting May 7, 2026.

The 15 June 2026 entry, headed "Removing documentation for the FAQ rich result feature":

What: Removed documentation for the FAQ rich result feature. Why: The FAQ rich result feature is no longer shown in Google Search results, as announced in the changelog entry in May 2026.

Two things follow from the wording. The feature death and the doc death are separate events five weeks apart, which is why you will find posts dated between them that describe the doc as still live. And the deprecation notice went up the day after the feature stopped, not before, so there was no warning window.

You can verify the end state without trusting anyone. Request developers.google.com/search/docs/appearance/structured-data/faqpage and watch where it lands. When we checked on 15 August 2026 it redirected to /search/updates#removing-faq-rich-result, the changelog anchor for its own removal.

This was not sudden so much as slow. Google had already narrowed FAQ results in August 2023, when a Search Central post said the result would "only be shown for well-known, authoritative government and health websites. For all other sites, this rich result will no longer be shown regularly." So for most sites the feature had been dead for nearly three years before it was officially buried. If your FAQ markup stopped producing anything visible in 2023, you already lived through the change that mattered.

What did not change

Four things, and they are the reason "just delete it" is bad advice.

The vocabulary. schema.org/FAQPage is a live type with mainEntity pointing at Question objects that carry acceptedAnswer. There is no deprecation flag on it. Google removing a search feature is Google editing Google's product, not schema.org editing the vocabulary.

Validity. Your markup does not become an error. Google was explicit about this situation in 2023, and the sentence has not been retracted: "While you can drop this structured data from your site, there's no need to proactively remove it. Structured data that's not being used does not cause problems for Search, but also has no visible effects in Google Search."

Your rankings. FAQ markup was never a ranking signal. It was a display feature, and losing a display feature changes your click-through rate on the queries that showed it, not your position.

The page itself. If your FAQ section is genuinely useful, it was useful for reasons that had nothing to do with the accordion. Retiring the markup does not retire the answers.

The gov.uk page still carrying FAQPage today

We ran gov.uk/vat-rates through our schema validator on 15 August 2026, three months after the feature died. The page returns two JSON-LD blocks. One is a complete FAQPage with two Question nodes, author, publisher, datePublished and dateModified. The other is a BreadcrumbList.

Trimmed to the shape:

{
  "@context": "http://schema.org",
  "@type": "FAQPage",
  "name": "VAT rates",
  "datePublished": "2011-09-12T16:43:03+01:00",
  "dateModified": "2014-12-12T14:55:23+00:00",
  "author": {
    "@type": "Organization",
    "name": "Government Digital Service",
    "url": "https://www.gov.uk/government/organisations/government-digital-service"
  },
  "mainEntity": [
    {
      "@type": "Question",
      "name": "VAT rates",
      "url": "https://www.gov.uk/vat-rates",
      "acceptedAnswer": {
        "@type": "Answer",
        "url": "https://www.gov.uk/vat-rates",
        "text": "The standard VAT rate is 20%"
      }
    }
  ]
}

GOV.UK was in the last category still eligible for the FAQ result before it went, so it is one of the few publishers that had something to lose. It has not removed the markup. That is not proof the markup does something. It is evidence that a large, well-resourced publishing team reached the same conclusion Google's guidance points at: leaving it costs nothing.

The breadcrumb block on the same page uses a shape you rarely see in examples. Instead of name on the ListItem, it puts a Thing inside item carrying @id and name. Google's breadcrumb documentation explicitly allows this: "If you're using a Thing with a name instead of a URL to specify item, then name is not required." Two valid shapes, and a validator that only knows one of them will report a false error on a government website.

FAQPage, QAPage and Education Q&A are three different things

Google retired one of these three. The other two are live features, and mixing them up costs real work.

Google kept the two where the content comes from users or classrooms, and dropped the one a site owner writes alone. If you run a forum, a support community or a product Q&A where customers actually ask things, QAPage is a live feature and worth implementing properly. If you wrote both halves yourself, it is not eligible. Rebadging your FAQ as a QAPage to chase the result is the kind of thing Google's spam guidance exists for.

Does the markup help AI answers?

For Google's own AI features, no. Google says so directly.

Google's guide to optimizing for generative AI features lists overfocusing on structured data as a mistake, in these words: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add. However, it's a good idea to continue using it as part of your overall SEO strategy, as it helps with being eligible for rich results on Google Search."

So for AI Overviews and AI Mode, Google has told you directly that FAQPage is not a lever. For other answer engines, no operator has published documentation saying they weight FAQPage markup, and no independent study we can find has measured it. We have not run that measurement ourselves, so we are not going to imply a result. Nobody has published this.

The duller thing is true instead. Your answers are readable whether or not they sit inside JSON-LD, as long as they are in the HTML. That is the part worth spending on. Making your site readable to AI agents covers the mechanics, and how AI Overviews choose citations covers what is actually known about selection.

Keep, remove, or replace

The types that still produce something are covered in the schema types that still earn something in 2026. If you need FAQPage output for a non-Google consumer, our FAQ schema generator is free and needs no signup, along with the rest of the tools.

What to do next

Do nothing to your existing FAQ markup. That is genuinely the correct action and it takes zero minutes.

Then spend the time you just saved on the thing FAQ markup was a proxy for. You wanted your answers to be the ones search shows. Check whether they are, by looking at whether AI Overviews name your page for the questions your buyers type. Our product, SEOBuilder, runs that scan against live AI answers and returns the citation URLs so you can see who was named instead. Seven engines, to be exact about it: Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, Gemini, Microsoft Copilot and Claude.

Start with how to measure AI visibility if you want the method before the product. Read answer engine optimization if you want to know which parts of the AEO pitch survive contact with Google's documentation.