> Google Killed the FAQ Rich Result, Don't Panic

On May 7, Google quietly added a deprecation notice to its FAQ structured data documentation. As of that date, FAQ rich results no longer render in Google Search. The search appearance filters and rich result reports come out in June. Search Console API support ends in August.

No blog post. No explanation. The only on-record line from Google is one sentence: "Unused structured data doesn't cause problems for Search."

Translation: leave the markup, lose the SERP feature.

That framing is exactly backwards, and we'll explain why.

the road to deprecation

• August 2023. Google restricted FAQ rich results to "well-known, authoritative government and health websites." Overnight, every plumber, every B2B SaaS, every recipe blog with valid Question/Answer schema lost the snippet.
• May 7, 2026. Even the gov + health carve-out is gone. The feature is dead.
• June 2026. Search Console FAQ filters and rich result reports go away.
• August 2026. The Search Console API stops returning FAQ data.

For most of the SEO world, the rich result already evaporated in 2023. What changed last week is the paperwork. The feature is now formally dead.

why we're telling you to keep the schema anyway

Because "no rich result" is not the same as "no value." Two reasons it still matters, and the second is the one Google won't tell you about.

1. LLMs read structured data

FAQPage JSON-LD is some of the cleanest, most machine-parseable QA content on the open web. Every mainEntity is a self-labelled question with a self-labelled answer. ChatGPT, Perplexity, Claude, Gemini, Bing Copilot -- every crawler-fed assistant -- ingests this kind of structured content greedily. When somebody asks an AI assistant a question your page already answers, the schema is what makes your answer trivially extractable instead of buried in a <div> a model has to guess at.

If you care about being cited by AI tools -- and if you don't, you should -- FAQ schema is still pulling weight. The visible SERP feature was always the smaller half of the prize.

2. Google still sees it. "Unused" is not "unread."

Google's own line is "unused structured data doesn't cause problems." That's a careful sentence. It says nothing about whether structured data is read, weighted, or fed into entity understanding, AI Overviews, or ranking-adjacent quality signals. We don't get to read Google's ranking source code, but we do know they keep ingesting structured data on every crawl.

The rich result is the part that disappeared. The data ingestion didn't. Treat FAQ schema as a ranking-adjacent signal until proven otherwise -- and given that the cost of leaving valid markup in place is zero, that's the cheapest hedge in technical SEO.

what to do this week

• Don't strip it out. Removing FAQPage schema is a one-way edit that costs you LLM extractability and possible ranking signal for zero search-rendering gain. There is no upside to deletion.
• Don't fake it where it doesn't fit. If the page isn't actually a FAQ, the markup was always a stretch and the rich-result incentive is gone. No reason to bolt it on now.
• Keep your Question/Answer markup valid. Same rules as before -- visible question, visible answer, no hidden content. Working examples and a free generator here: FAQ JSON-LD snippet -- generator and real examples.
• Export your Search Console data before June. If you have a year of FAQ rich result impression history and want to keep it, pull it now. The report goes dark in roughly four weeks.

the bigger pattern

Google has been thinning out rich result types for two years. HowTo went. FAQ got restricted, then deprecated. Others have quietly migrated into AI Overviews instead of rendering as discrete features. The direction of travel is clear: the structured-data layer is something Google increasingly ingests rather than renders. The same is true on the LLM side. Markup originally written for the SERP is now load-bearing for everything that comes after the SERP.

FAQ schema being "deprecated" by Google is a SERP event, not a schema event. Schema.org didn't change. The crawlers didn't change. The downstream consumers of structured data -- search engines, AI assistants, voice assistants, every SaaS that scrapes the web for grounding data -- still parse it the same way they always did.

Leave the JSON-LD in. Mark up new pages that need it with valid FAQPage. Don't chase whatever the next deprecation announcement is. The schema layer was never about the snippet -- the snippet was always the bonus.