Add an FAQ section

What the scan checks: it counts your question-and-answer pairs, from either FAQPage schema or question-shaped headings on the page, and wants at least four. Content pages with fewer fail. (Homepages are exempt.)

A pair needs an answer. A question-shaped heading only counts when at least 20 words of answer follow it before the next heading. This used to count question headings alone, which meant a page of questions with nothing underneath them scored as an FAQ. Accordions (<details>/<summary>), definition lists and per-item wrapper markup all count normally.

A note on Google, so you do not delete your FAQ markup by mistake. Google withdrew FAQ rich results for most sites in August 2023 (government and health sites kept them) and retired the last of the associated Search Console reporting in May 2026. If you shipped FAQPage schema for the SERP decoration, that reason is gone.

The reason this check exists is different and is unaffected. Verbatim question-and-answer pairs are the shape answer engines lift most readily, and the structured-data analyses put the AI-citation lift for FAQ content at 20–44%. The value was never the rich result.

Why it matters for AI search visibility

An FAQ block is verbatim question-and-answer text, which is exactly what AI answer engines lift into responses. Google retired FAQ rich results for most sites in 2023, but for AI citations the pattern is very much alive: one 2026 structured-data analysis found a 20 to 44% AI citation-rate lift for FAQPage markup, and layering it with other complementary schema types roughly doubled citation frequency. Each Q&A pair is a pre-segmented, self-contained chunk built around a real query, which is the ideal unit for retrieval. It is one of the highest-return content additions available.

How to fix it

Add a real FAQ section with at least four pairs:

• Use genuine user questions. Mine your search queries, sales calls, and support tickets. “Does it work with Shopify?” beats an invented question nobody asks.
• Answer in 40 to 60 words, self-contained, so each answer stands alone as a chunk.
• Put the questions in <h2>, <h3>, <dt>, or <summary> elements so they are detectable as questions, and make sure the text is in the raw HTML (accordions that lazy-load answers on click are invisible to bots; hide with CSS instead).

Then pair the visible FAQ with FAQPage JSON-LD so engines get the structured version too:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "Does it work with Shopify?",
    "acceptedAnswer": { "@type": "Answer", "text": "Yes. The scanner reads any live URL..." }
  }]
}

The schema text must match the visible text exactly. See C17.