If you do one thing for AEO this week, do this. FAQ schema is the single highest-impact piece of structured data you can add to a website — and most sites don't have it. Here's exactly how to add it, with a paste-ready example.
FAQ schema is a small block of JSON-LD code you add to a webpage that tells search engines, "This page contains questions and direct answers." Specifically, it uses Schema.org's FAQPage type to map each question (Question) to its accepted answer (Answer). Search engines and AI engines parse it the same way — by extracting a clean question-answer pair they can quote in a response.
You don't need to redesign your page or rewrite content. You add a single <script type="application/ld+json"> tag in the <head> of any page that has FAQ-style content. That's it.
AI search engines like ChatGPT, Perplexity, Google AI Overviews, and Bing Copilot don't read your page the way a human does. They parse the HTML and look for clean, extractable facts. A page with structured FAQ schema is delivering those facts on a silver platter.
Research from Previsible across 5,000 AI search prompts found that pages with question-answer formatting get cited 180% more often than equivalent prose content (Previsible, 2024). And from the Cited Digital audit corpus across 1,168 sites, FAQPage schema is missing on 89% of sites — including most well-ranked SEO sites. It's the single biggest underused signal.
FAQ content is roughly 3x more powerful for AI citation than FAQ schema alone. The combination of both — visible question-answer text plus matching JSON-LD schema — is the maximum signal you can give an AI engine. (SE Ranking, 2024)
Almost every business site benefits from it, but FAQ schema is especially high-impact for:
If your customers ask questions before they buy — and they do — FAQ schema turns the answers you've already written into AI-extractable assets.
Open ChatGPT or Perplexity, type "common questions about [your category]" — say, "common questions about hiring a roofer" — and write down the top 8 to 12 questions you see. These are the queries AI engines are already answering, with or without you. They're your target list.
Cross-reference with: questions you actually answer in customer emails, search-console long-tail queries that include question words, and the People Also Ask section on Google for your top keywords.
Each answer should be 2–4 sentences, plain English, no jargon. AI engines extract verbatim, so write the way you'd want to be quoted. Include specific numbers, dates, and named entities where possible — they boost extractability.
The schema works best when it matches visible content on the page. Use real <h2> or <h3> tags for each question, with a paragraph (or bullet list) underneath for the answer. Don't hide the content in a JavaScript-rendered modal — AI crawlers often don't execute JavaScript, so they'd miss it.
You can use HTML <details>/<summary> elements for an accordion UI; they're crawler-friendly because the content is in the static HTML.
Paste this template into the <head> of your page. Replace the question and answer text with yours. Add or remove Question blocks as needed.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "How much does a roof replacement cost in 2026?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Asphalt-shingle roof replacement on a typical 1,800 sq ft home runs $7,000-$12,000 in 2026. Material grade and tear-off complexity drive most of the variance."
}
},
{
"@type": "Question",
"name": "How long does a roof replacement take?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Most asphalt-shingle replacements take 1-2 days for a single-family home. Metal and tile roofs can take 3-5 days. Weather, deck repairs, and chimney flashing add time."
}
}
]
}
</script>Paste your page URL into Google's Rich Results Test. It will parse your JSON-LD, flag any errors, and confirm the FAQ markup is detected. Fix anything red.
Also worth running: Schema.org's validator — it's stricter and catches edge cases the Google tool misses.
AI engines need 2–4 weeks to recrawl, parse, and start citing schema-marked content. After that window, you can test directly: ask ChatGPT or Perplexity a question your FAQ answers, and see if your page appears as a citation.
| Mistake | Why it fails |
|---|---|
| Schema text doesn't match visible page content | Google flags this as deceptive markup and ignores the schema entirely. |
| Including marketing copy or CTAs in the answer | "Call us today!" inside an Answer block reads as low-quality content. AI engines deprioritize it. |
| Only one or two questions | Single-question FAQ markup gets less weight. Aim for 5–12 per page. |
| Same FAQ block on every page | Boilerplate FAQ duplicated site-wide gets devalued. Tailor questions to each page's topic. |
| FAQ rendered only by JavaScript | If the questions/answers aren't in the static HTML, most AI crawlers won't see them. |
| Mixing Question/Answer with HowTo/Step | Different schema types — pick one per content block. Use HowTo for ordered steps, FAQPage for Q&A. |
Indirectly — yes. FAQ schema doesn't directly boost your blue-link ranking, but it dramatically improves your odds of being extracted into Google AI Overviews, the People Also Ask box, and rich snippets. All three steal clicks from page-1 results, so being cited in them is more valuable than ranking #2 below them.
Aim for 5–12. Below five, the markup is treated as light. Above twelve, you're padding and the answers tend to weaken. Quality over quantity.
Yes — and you should. Each page's FAQ should be tailored to that page's topic. A pricing page's FAQ answers pricing questions. A service page's FAQ answers questions about that service. Don't reuse the same FAQ block across the site.
Yes, if the schema and the matching visible content are server-side rendered or pre-rendered. If both are injected at runtime by JavaScript, you're at the mercy of each crawler's JS-rendering behavior. Google handles it; ChatGPT and Perplexity often don't.
Schema typically shows up in Google search experience within 2–4 weeks. AI Overviews and ChatGPT citation can take 4–12 weeks because those systems update their knowledge graphs less frequently. Be patient and don't churn the schema.
Then add one. Even a 3-question FAQ (with thoughtful, specific answers) on a service page is better than no schema at all. Pick the three questions every customer asks before buying. Write the answers. You're done.
FAQ schema is the single most underutilized AI-search signal across the small-business web. It takes ten minutes per page to add. It produces measurable lift in AI citation rates within a month. And in our 1,168-site audit corpus, almost nine out of ten sites are missing it.
If you've read this far and your site doesn't have FAQPage schema, that's the next thing on your list. Start with your highest-traffic page. Pick five questions. Write the answers. Paste the JSON-LD. Validate. Move on to the next page.
Questions? Contact Connor