We disagree with the consensus on FAQ schema — here's our actual position

The standard advice in the structured data community is that FAQ schema is one of the highest-leverage additions you can make for AI visibility. We've seen this repeated in guides, agency pitches, and forum posts. We don't think it's right, and we want to explain why. **The case for FAQ schema** The argument goes: AI systems that answer questions draw on FAQ-structured content because it's already in question-and-answer format. Adding FAQ schema makes your content easier for AI systems to parse and cite. This is plausible, and there's some evidence that FAQ schema improves featured snippet performance in traditional Google search. **Why we're sceptical for AI citations specifically** The problem is that AI systems like ChatGPT, Perplexity, and Claude are not primarily drawing on structured data when they answer questions. They're drawing on the quality and authority of the underlying content. FAQ schema is a wrapper around content — it tells crawlers "this section is structured as questions and answers." But if the underlying content is thin, repetitive, or low-quality, the schema wrapper doesn't rescue it. The businesses we expect to be cited most by AI systems are the ones with the most authoritative content, not the most schema markup. We'll be tracking this as part of our six-month data project (see the Share Your Results thread — we've committed to publishing the actual findings at the six-month mark). **What FAQ schema is actually good for** It's genuinely useful for Google's featured snippets and "People also ask" boxes. If you're optimising for traditional search, FAQ schema is worth adding. If you're optimising specifically for AI citation volume, we'd prioritise Organisation schema, `sameAs` links, and content quality over FAQ schema. **The honest caveat** We're working from a short time window and a developing landscape. The AI search environment is changing fast, and what's true today may not be true in six months. We'll track this as part of our six-month data project. If the data shows we're wrong, we'll say so. But right now, if an agency is telling you that FAQ schema is the primary lever for AI visibility, we'd push back on that framing.