In short
FAQPage schema marks up genuine question-and-answer pairs already on your page so both search engines and AI systems can parse them as clean, citable units. Write real answers, keep them self-contained, and only mark up content a reader can actually see.
What FAQ schema is for
FAQPage is a JSON-LD schema type that tells a search engine "here are question-and-answer pairs, verbatim, on this page." It was built for search rich results, but it turns out to be exactly the shape an AI answer engine wants too: a question paired with a self-contained answer is the easiest thing in the world to lift and cite.
The one rule that matters most
Only mark up questions and answers that are actually visible on the page in the same words. Structured data is a description of your content, not a way to sneak in extra content invisibly — search engines and AI crawlers both treat mismatches between markup and visible text as a trust problem, and it can get the markup ignored or the page penalized.
The format
A minimal, valid FAQPage entry looks like this:
``json { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [{ "@type": "Question", "name": "How long does a rebrand take?", "acceptedAnswer": { "@type": "Answer", "text": "A typical rebrand takes six to ten weeks, from discovery through final delivery." } }] } ``
Repeat the Question/Answer pair for every FAQ on the page, inside one mainEntity array, inside one <script type="application/ld+json"> tag.
Writing answers that get quoted
- Answer in the first sentence. Don't build up to it — state the fact, then explain.
- Keep it self-contained. The answer should make sense with zero page context, since that's exactly how it'll be lifted and shown.
- Use real numbers and specifics over vague qualifiers. "Six to ten weeks" beats "it depends on scope."
- One idea per answer. If an answer needs two paragraphs to cover two different things, split it into two questions.
Where FAQ schema fits in a page
It works best paired with actual FAQ content the reader can see — a "Common questions" section near the end of an article or service page, not a hidden accordion stuffed with extra keywords. If you're building out full articles, an Article + BreadcrumbList + FAQPage combination (all three together) is the strongest baseline — see the complete guide to GEO for how the pieces fit together.
Validating it
Before shipping, run the page through Google's Rich Results Test and check for two things: zero errors, and that every question/answer pair in the JSON-LD matches the visible text word-for-word. A common bug is editing the visible FAQ copy later and forgetting to update the matching JSON-LD — schedule a periodic check if your FAQs change often.
What it actually gets you
Done right, FAQ schema can earn expandable rich results in search (more visual space, higher perceived authority) and makes your content dramatically easier for an AI system to quote accurately, because you've done the parsing work for it in advance.
Common questions
Can I mark up FAQ schema on any page?
You can technically add it anywhere, but it should only go on pages that contain a genuine, visible FAQ section — using it as a workaround to add extra text is against the intent of the format and can be treated as spam.
How many questions should I include?
As many as are genuinely useful — three is fine, twenty is fine, as long as each one answers something a real visitor would ask.
Does FAQ schema help AI citations specifically?
Yes — the question/answer format is naturally easy for a model to lift as a self-contained, quotable unit, which is part of why it's covered in our how to get cited by AI guide too.