FAQPage Schema
FAQPage schema is the Schema.org type that declares a page contains a list of frequently asked questions with answers. Historically powered Google's visual FAQ rich snippets (removed May 7, 2026) but remains critical for AI Mode extraction, voice-assistant answers, and ChatGPT/Perplexity/Claude RAG citation. Still worth implementing despite the May 2026 visual change.
Also called: FAQPage JSON-LD, FAQ schema · Last updated: May 27, 2026 · By Joseph W. Anady
Why it matters.
FAQPage schema is a single page-level declaration containing an array of Question entities, each with an acceptedAnswer. The structure mirrors how users actually search in voice and conversational interfaces. Even after the May 7, 2026 visual change removed FAQ accordions from SERPs, FAQPage schema continues to be extracted by AI engines for citation — the visual reward disappeared but the structured-extraction reward remains.
How it works.
FAQPage schema lives in a JSON-LD block on the page. Required: @context, @type FAQPage, mainEntity (array of Question objects). Each Question requires name and acceptedAnswer with @type Answer and text. The text in acceptedAnswer is the extractable answer — should be self-contained 40-100 words that don't require surrounding context.
2026 reality check.
May 7, 2026: Google removed FAQ rich snippets from SERPs. August 2026: removed FAQPage reporting from Search Console. Many practitioners interpreted this as 'FAQ schema is dead.' Wrong. FAQPage schema is still extracted by AI Mode, ChatGPT, Perplexity, and Claude for response generation. The visual SERP feature died; the structured-data extraction lives. Keep the schema.
Data points
- Schema.org FAQPage documented at https://schema.org/FAQPage
- Visual FAQ rich snippets removed May 7, 2026
- Search Console FAQPage reporting removed August 2026
- Still extracted by AI Mode, ChatGPT, Perplexity, Claude (multiple 2026 confirmations)
- Each answer should be 40-100 words self-contained (extractable atom)
First-hand insight from ThatDeveloperGuy.
ThatDeveloperGuy ships FAQPage schema on every content page across the network. Despite the May 2026 visual change, AI Mode citations on TDG-content queries continued at the same rate post-change. Our internal hypothesis: FAQPage structured data is still the cleanest extraction signal for AI engines, even though Google no longer displays the visual accordion. We maintain it on every page.
How TDG approaches it
Every TDG content page has 5-8 FAQ Q&A with conversational questions matching voice and AI query patterns. Answers are 40-100 words, self-contained, no marketing prose. The visible HTML mirrors the FAQPage schema (Google parses both). Maintained even on every page across the network.
Common mistakes.
- Removing FAQPage schema after the May 2026 visual change (loses AI citation eligibility)
- Writing answers that require surrounding context (answers must be self-contained)
- Stuffing the schema with marketing copy instead of direct factual answers
- Skipping question-form headers in the visible content (schema should mirror visible Q&A)
- Too many questions (focus 5-8 high-value Qs, not 50)
FAQ.
Is FAQ schema dead after May 2026?
No. The visual FAQ accordion was removed from SERPs on May 7, 2026 (Google says it produced overly busy results). But FAQ schema is still extracted by AI Mode, ChatGPT, Perplexity, and Claude for response generation. The reward shifted from visible SERP feature to structured AI extraction.
Why did Google remove FAQ rich snippets?
Officially: to simplify the SERP experience. Practically: FAQ accordions were heavily over-used and Google decided the visual clutter outweighed the user benefit.
How many FAQs per page?
5-8 is the sweet spot. Each question should be high-buyer-intent (something someone would actually search). 50+ FAQs dilutes signal.
Can I use FAQPage on every page?
Yes if the page genuinely contains a Q&A section. Don't fabricate FAQs to gain schema — Google's spam filters catch this.
Should the visible FAQ match the schema exactly?
Yes. Schema-visible content mismatch is a quality signal Google penalizes. The HTML FAQ should mirror the JSON-LD FAQ.
Maintained by Joseph W. Anady at ThatDeveloperGuy. Back to glossary · Suggest a term