In August 2023, Google made a change that quietly damaged how millions of businesses approach their websites. They restricted FAQ rich results, the expandable question-and-answer accordions that appeared under search results, to authoritative government and health websites. Everyone else lost them overnight. The SEO industry interpreted this as a signal: FAQ schema is dead. Agencies told clients to stop implementing it. Audit tools stopped flagging it as an opportunity. And over the next three years, businesses across every industry quietly dropped FAQPage schema from their sites.
Our audit of 67 restaurant websites confirms the damage. Only 1 out of 67 has FAQPage schema. That is not because restaurants do not have FAQ content. It is because someone told them FAQs do not matter anymore. That advice was wrong, and it is costing them visibility in AI search.
What Google actually said vs what the industry heard
Google made a factual announcement: FAQ rich results would be restricted to government and health websites. Google did not say to stop using FAQ schema. They did not say FAQ pages no longer matter. They removed a visual SERP feature, the accordion widget, while leaving the underlying structured data fully supported and indexed.
But as the message traveled from Google to SEO press to agencies to website owners, each layer simplified it. Google reduced a SERP feature became FAQ schema is worthless. Here is how the chain reaction worked:
- Google Search Liaison Danny Sullivan announced the restriction on FAQ rich results. Factual, limited in scope.
- SEJ, Search Engine Land, and Lily Ray at Amsive published tracking data showing massive FAQ snippet losses. Accurate coverage, but the framing emphasized the loss, not the nuance.
- SEO audit tools like Ahrefs and Semrush stopped flagging missing FAQPage schema as an opportunity because the SERP feature was gone.
- Agencies and consultants told clients not to waste time on FAQ schema. The reasoning: if Google does not reward it visibly, why bother.
- Website owners and CMS templates followed suit. FAQ sections were removed. FAQPage schema was stripped from boilerplate templates.
What got lost: Google removed the accordion. They did not remove indexing of FAQ content. And critically, AI search engines like ChatGPT, Perplexity, and Claude never used the Google SERP accordion. They read the schema directly from your HTML. The deprecation that killed a Google feature made the same schema MORE valuable for a faster-growing channel.
Why FAQPage schema matters more now than in 2023
When ChatGPT, Perplexity, or Google AI Overviews generates an answer, it looks for authoritative sources to cite. FAQPage schema wraps content in a format the AI can parse with near-perfect accuracy. Every Question and Answer pair is a clean, machine-readable unit. Compare that to a paragraph of prose that the AI has to interpret and guess at. Schema removes the guessing.
Think about how people actually use AI search. They ask questions. Does this restaurant have vegetarian options? What time do they open on Sunday? Do they take reservations? Is there parking? Each of those is a natural-language question that maps directly to a FAQPage Question/Answer pair. If your site has that schema, you are the citation source. If it does not, the AI guesses from whatever unstructured text it can find, if it finds anything at all.
In our 67-site audit, the one restaurant with FAQPage schema (Zen Japanese, eatzen.com) is essentially the only restaurant in the entire sample positioned to be cited by AI for question-based queries. Not because their food is better or their brand is bigger. Because they have the schema and nobody else does.
The three layers of AI search visibility
To understand why FAQ schema matters and how small businesses can compete with big brands, you need to see the full puzzle. AI search visibility is not one thing. It is three layers stacked on top of each other, and most of the SEO industry only talks about the first one.
Layer 1 is Technical: Can AI read your site? This is what our audit measures. Schema, crawlability, AI bot access, sitemap, server-side rendering. Chains fail this layer badly because their JavaScript sites return empty HTML. Independents pass it because their simpler sites produce crawlable HTML by default. This layer is necessary but not sufficient. You cannot win on Layer 1 alone.
Layer 2 is Entity: Does AI know who you are? Google and AI models understand the world through entities. McDonald's is a strong entity. It exists in Wikipedia, Google's Knowledge Graph, thousands of news articles, and the training data of every major LLM. When ChatGPT thinks McDonald's, it pulls from a rich web of corroborating data. An independent restaurant is a weak entity. It might exist in Google Business Profile and Yelp, but it has thin corroboration. The AI is not sure what it is. This is where chains have their real advantage. Not in their website, but in their entity strength.
Layer 3 is Authority: Does AI choose you? When the AI answers a question, it picks from candidates it knows about and trusts. Authority signals, links, mentions, citations across the web, determine who gets chosen. A chain with millions of backlinks wins here. An independent with 30 backlinks does not.
The chain advantage is in Layers 2 and 3. The independent advantage is in Layer 1. The winning strategy for small businesses is to win Layer 1 so decisively that it compensates for Layer 2 weakness, while simultaneously building entity strength through structured data. Structured data IS entity building. When you add Restaurant schema with your name, address, cuisine, and hours, and that same information appears consistently on your Google Business Profile, Yelp listing, and llms.txt, you are telling every AI system that you are a real, specific, verifiable entity. You are manufacturing the corroboration that chains get for free.
How small brands compete with big brands
Chains can coast on brand. McDonald's scores 0/100 on our audit and it barely matters today because people type McDonald's directly. They do not search for burger restaurant near me and hope McDonald's shows up. The AI models have McDonald's in their training data. Google Business Profile carries location queries. Third-party platforms like Yelp and DoorDash serve as proxy content. McDonald's does not need FAQPage schema because its brand is so embedded that AI can describe it from background knowledge.
But that brand advantage only works for branded queries. It breaks down completely for category-based and conversational queries. When someone asks ChatGPT where can I get a gluten-free burger in Austin, the AI does not default to McDonald's. It looks for the most legible, well-structured answer it can find. A small restaurant with FAQPage schema answering that exact question is positioned to be cited. McDonald's, with its empty JavaScript site and no schema, is not.
- Branded queries like 'McDonald's menu': Brand carries you. Schema does not matter.
- Category queries like 'burger restaurant near me': Brand helps partially. Schema starts to matter.
- Conversational queries like 'Where can I get a gluten-free burger in Austin?': Brand does not help. Schema is critical.
- Long-tail queries like 'What restaurants are open late on Sunday with vegetarian options?': Brand does not help. Schema is critical.
The shift from keyword search to conversational AI search is moving everyone toward the bottom of that list. And in that world, the chain's biggest advantage, brand recognition, becomes irrelevant for the queries that are growing fastest.
Who is using FAQPage schema and winning right now
I am going to be honest here rather than make claims I cannot back up. I cannot point to a specific brand and say this company is dominating AI search because of FAQPage schema. That kind of attribution data does not exist yet because AI search engines do not publish which schema types get cited most. What I can point to is logic and evidence.
On Google's traditional SERP, health and government sites kept their FAQ rich results after the 2023 change. WebMD, Healthline, Mayo Clinic, and the NHS are the sites Google explicitly exempted. They still get the accordion feature. Whether that translates to measurable business advantage is debatable, but they have the visibility and nobody else does.
On AI search, the picture is murkier but the logic is clear. FAQPage is the cleanest machine-readable Q&A format. AI systems that parse structured data will preferentially use it over unstructured text. Our audit proves the gap exists: 1 out of 67 restaurants has it. That is an opportunity, not proof that anyone is currently capitalizing on it. The businesses that add it now, when 98.5 percent of their competitors do not have it, will be the ones positioned to capture that advantage.
What to do right now
If you are a small or independent business, here is the strategic reality. You cannot out-brand a chain. You cannot outspend them on Layer 3 authority building. But you can win Layer 1 technical visibility so decisively that it compensates for entity weakness, and you can build entity strength through structured data and consistency across the web. Here is the priority list:
- Add FAQPage schema answering the 5 to 10 most common questions customers ask about your business. This is the single highest-leverage thing you can do for AI citation visibility.
- Add Restaurant or LocalBusiness schema with your full NAP (name, address, phone), hours, cuisine type, and price range. This builds your entity identity.
- Make sure your NAP is identical across your website, Google Business Profile, Yelp, and any other directory. Consistency is what turns scattered listings into a strong entity.
- Allow AI crawlers. Check robots.txt for Disallow rules on GPTBot, ClaudeBot, PerplexityBot, and Google-Extended.
- Create an llms.txt file that tells AI crawlers what your business is and which pages matter most.
None of this requires a developer for most sites. WordPress, Squarespace, and Wix all support FAQPage schema through built-in tools or free plugins. The total investment is 2 to 4 hours. The competitive window is open right now because 98.5 percent of your competitors do not have it. It will not stay open forever.
Parceit audits your site for FAQPage schema, entity data, AI crawler access, and 20+ other signals in under 60 seconds. See your score free at parceit.com/audit. If you want it implemented for you, we fix everything for $399 flat.
Frequently asked questions
Did Google kill FAQ rich results?
Google restricted FAQ rich results (the expandable accordion in search results) to government and health websites in August 2023. This removed the visual SERP feature for most sites but did not stop Google from indexing FAQ content or FAQPage schema. The schema is still fully supported by Google and is now more important because AI search engines use it directly.
Should I still use FAQPage schema in 2026?
Yes. FAQPage schema is the single most important structured data type for AI search citations. AI engines like ChatGPT, Perplexity, and Google AI Overviews parse it to extract question-answer pairs for their answers. Our audit of 67 restaurant websites found only 1 with FAQPage schema. The gap between its importance and its adoption is the biggest opportunity in search visibility right now.
How do small businesses compete with big brands in AI search?
Big brands win on entity recognition (Layer 2) and authority (Layer 3) because they have Wikipedia pages, news coverage, and millions of backlinks. Small businesses win on technical visibility (Layer 1) because their simpler websites are more crawlable by AI. The strategy is to win Layer 1 so decisively that it compensates for entity weakness, while using structured data to build entity strength over time. For conversational queries like 'Where can I get a gluten-free burger in Austin?', brand advantage disappears and schema quality determines who gets cited.
Who pushed the idea that FAQ schema is dead?
No single entity killed it. Google announced the FAQ rich result restriction in August 2023. SEO press (Search Engine Journal, Search Engine Land, Search Engine Roundtable) covered it accurately but framed it as a loss. Lily Ray at Amsive Digital published tracking data showing massive FAQ snippet losses. SEO audit tools like Ahrefs and Semrush stopped flagging missing FAQPage schema as an opportunity. Agencies told clients not to bother. Each layer simplified the message until 'Google reduced a SERP feature' became 'FAQ schema is worthless.'