Rank Math is one of the most popular WordPress SEO plugins available, and its FAQ block feature is genuinely useful for adding structured data to your pages without writing a single line of code. But if your goal is to appear in AI-generated answers on ChatGPT, Perplexity, or Gemini, the question of whether Rank Math's FAQ schema actually does the job is worth examining closely.
The short answer: it partly works, but with meaningful limitations that most site owners don't know about. Let's break it down properly.
What Rank Math's FAQ Block Actually Outputs
When you use Rank Math's FAQ block inside the WordPress block editor (Gutenberg), it automatically generates a FAQPage schema object in JSON-LD format and injects it into the page's <head>. This is the correct approach. JSON-LD in the head is exactly what Google's documentation recommends, and it's also the format most likely to be parsed cleanly by AI crawlers.
The output typically looks like this:
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "Your question here",
"acceptedAnswer": {
"@type": "Answer",
"text": "Your answer here"
}
}
]
}That's a well-formed, valid schema structure. It passes Google's Rich Results Test without issues. For traditional SEO purposes, specifically triggering FAQ rich results in Google Search, it works well enough.
But AI search engines are not Google. They don't index pages the same way, they don't evaluate structured data the same way, and they don't reward the same things.
How ChatGPT and Perplexity Actually Use Structured Data
This is where things get more interesting. ChatGPT (particularly when using its browsing mode) and Perplexity both retrieve content from the web in real time to answer queries. They're looking for pages that clearly, confidently answer specific questions. Schema markup is one signal they can use, but it's not the only one, and it's arguably not even the primary one.
Here's what actually matters for AI citation:
- Answer completeness. The answer text inside your
acceptedAnswerfield needs to be substantive. If your FAQ answers are one sentence long, they're unlikely to satisfy an AI looking to give a thorough response to a user query. - Topical authority signals. AI engines look at the broader page context, not just the schema block. A strong FAQ block on a thin page is not going to outperform a well-written article that happens to include structured FAQ data.
- Crawlability and freshness. Perplexity in particular relies on recent, crawlable content. If your site has poor crawl health, even perfect schema won't help.
- Entity clarity. Who is answering these questions? If your page lacks clear authorship, brand identity, or Organisation schema, the AI has less context for why it should trust your answer over someone else's.
Rank Math's FAQ block handles the technical formatting correctly. It does not handle any of the above.
The Specific Gaps in Rank Math's FAQ Schema Output
Answer text is often truncated or stripped
One issue that comes up regularly is that Rank Math's FAQ block strips HTML from the answer text in the JSON-LD output. If your visible answer on the page includes bold text, links, or lists, the schema output may only contain a plain-text version. This isn't necessarily fatal, but it does mean the structured data and the visible content can diverge. AI engines that compare the two may deprioritise the page as a result.
No connection to broader page schema
Rank Math generates the FAQPage type as a standalone entity. It's not nested within or connected to your Article, Product, or WebPage schema. For traditional Google SEO this is fine, but for AI engines trying to understand the full context of a page, disconnected schema objects tell an incomplete story.
A well-structured page for AI visibility would typically have a parent entity (say, an Article or WebPage) with the FAQ data referenced as part of the page's overall content structure. Rank Math doesn't do this automatically.
No control over @id or entity relationships
Schema works best for AI when entities are linked. Your FAQ page ideally references your Organisation, your author (via Person schema), and your broader content graph. Rank Math gives you very limited control over @id values and entity linking within its FAQ block. Advanced users can sometimes work around this by editing schema in Rank Math's Schema Generator, but it requires technical knowledge most users don't have.
Duplicate FAQPage types can cause conflicts
If you use multiple FAQ blocks on a single page, Rank Math can output multiple FAQPage objects. In some configurations this causes validation errors. It can also confuse AI parsers that expect a single, coherent FAQPage entity per URL.
Where Rank Math's FAQ Schema Does Work Well
To be fair, there are scenarios where Rank Math's FAQ schema genuinely contributes to AI visibility.
If you're writing well-structured, long-form content and adding FAQ blocks that contain thorough, specific answers, the schema reinforces what's already a strong page. The structured data isn't the reason the AI cites you. It's the combination of good content, proper schema, and strong topical authority. Rank Math's output is good enough to not be the weak link in that chain.
It also handles the basics reliably. The JSON-LD is valid, it's in the right location, and it uses the correct schema.org types. That matters. Plenty of sites have broken FAQ schema that actively hurts their chances of being parsed correctly.
For smaller WordPress sites that don't have the budget for custom schema implementation, Rank Math's FAQ block is a solid starting point. Just don't treat it as a complete solution for AI visibility.
How to Strengthen Rank Math's FAQ Schema for AI Citations
If you're currently using Rank Math and want to improve your chances of being cited by ChatGPT or Perplexity, here are practical steps you can take without abandoning the plugin.
Write longer, self-contained answers
Each answer in your FAQ block should be able to stand alone. Aim for at least 60 to 100 words per answer. Think about what a user would need to read if this answer appeared in a Perplexity citation box with no other context. Would it make sense? Would it be genuinely useful? If not, expand it.
Add Organisation and Author schema separately
Rank Math does support Organisation and Person schema through its other schema types. Make sure your homepage has a properly configured Organization schema with your name, logo, URL, and social profiles. Add Person schema to your author pages. This gives AI engines the entity context they need to trust your FAQ content.
There's a detailed walkthrough on how to use Person schema to build authority in AI search that covers this side of things well.
Complement the block with manual JSON-LD where needed
For your most important pages, consider supplementing Rank Math's output with hand-written or custom JSON-LD that connects your FAQ data to the broader page entity. You can inject additional JSON-LD without conflicting with Rank Math's output, as long as you're careful not to duplicate types. The guide on adding JSON-LD to WordPress without a plugin explains how to do this safely.
Test in Google's Rich Results Test and review the raw JSON
Always check what Rank Math is actually outputting. The Rich Results Test shows you the parsed schema, and you can also view the raw JSON-LD in your page source. Look for truncated answer text, duplicate FAQPage objects, or missing fields. Fix issues before they affect your AI visibility.
Rank Math vs Custom Schema for Serious AI Visibility
There's a point at which relying on plugin-generated schema becomes a limiting factor. Rank Math is built to serve a wide range of WordPress sites with varying needs. It's not optimised specifically for AI search visibility, and it probably never will be, because that's not its primary purpose.
If you're an e-commerce brand or content business that depends on appearing in AI-generated recommendations and answers, the gap between "valid plugin schema" and "AI-optimised structured data" is real and growing. AI engines are getting better at understanding entity relationships, author authority, and content depth. Schema that just checks the basic validation boxes is increasingly not enough.
This is the kind of problem FlinnSchema works on directly. Rather than patching plugin output, the approach involves building a proper schema architecture for the whole site, with entity relationships, connected knowledge graphs, and FAQ data that's genuinely aligned with how AI engines retrieve and cite content. You can request a free AI visibility audit to see exactly where your current setup stands.
The comparison between Rank Math, Yoast, and Schema Pro for AI visibility is worth reading if you're evaluating your options. Take a look at the breakdown of which WordPress schema plugin is best for AI visibility for a side-by-side view.
Frequently Asked Questions
Does Rank Math's FAQ schema get you into Google's FAQ rich results?
Yes, in most cases. Rank Math outputs valid FAQPage JSON-LD that meets Google's requirements for FAQ rich results. Whether you actually appear depends on other factors like your site's quality signals, but the schema itself is usually sufficient for Google's purposes.
Does Perplexity read FAQ schema directly?
Perplexity crawls pages and reads both the visible content and structured data. It doesn't rely on schema alone to decide what to cite. Strong FAQ schema on a well-written page increases the clarity of your content for Perplexity's retrieval process, but schema without good underlying content is unlikely to drive citations on its own.
Can I use Rank Math and also add custom JSON-LD on the same page?
Yes. Multiple JSON-LD blocks on a single page are valid as long as they don't duplicate the same schema type for the same entity. You can add a custom JSON-LD script tag via a WordPress hook or a code block in the editor, and it will work alongside Rank Math's output without conflict in most cases.
What's the biggest thing Rank Math's FAQ schema is missing for AI search?
Entity linking. Rank Math generates a valid but isolated FAQPage object. It doesn't connect your FAQ content to your Organisation, your author identity, or the broader page entity graph. AI search engines are increasingly using entity relationships to evaluate trustworthiness and relevance, and that's the gap that plugin-generated schema struggles to fill without additional customisation.