We Audited 67 Restaurant Websites for AI Search Visibility. The Results Are Embarrassing.

When someone asks ChatGPT where to eat, does your restaurant show up? We decided to find out by running our audit engine against 67 real restaurant websites. We crawled 18 major chains (McDonald's, Starbucks, Subway, Chipotle, Olive Garden, and 13 more) and 49 independent restaurants and cafes across Portland, Austin, and Asheville. The results reveal an industry that is fundamentally unprepared for AI search, and they contain one genuinely surprising finding: independent restaurants are beating the chains.

This is not opinion or speculation. Every score in this article comes from a real crawl of the live website, measuring the same signals that GPTBot, ClaudeBot, PerplexityBot, and Google's AI crawlers use to discover and understand content. You can verify any of these results by running the same audit on any URL at parceit.com.

The headline numbers

Across all 67 sites, the picture is bleak. The average score was 38 out of 100. Eleven sites scored absolute zero, meaning AI crawlers see nothing at all. Only one site out of 67 had FAQPage schema, the single most important structured data type for getting cited in AI answers. Here is the full breakdown:

Average score across all 67 sites: 38/100
Sites scoring zero (invisible to AI): 11 out of 67 (16%)
Sites with FAQPage schema: 1 out of 67 (1.5%)
Sites with LocalBusiness or Restaurant schema: 19 out of 67 (28%)
Sites with llms.txt: 18 out of 67 (27%)
Sites allowing all 4 major AI crawlers: 51 out of 67 (76%)

The single most alarming statistic: only 1 restaurant out of 67 has FAQPage schema. When ChatGPT or Perplexity answers a question like 'Does Kerbey Lane have gluten-free options?' or 'What time does Stumptown open?', FAQPage is the structured data format that makes your answer citable. Without it, the AI guesses from whatever unstructured text it can find, if it finds anything at all.

Chains vs independents: the surprise

We expected the big chains, with their enterprise CMS platforms and six-figure web budgets, to crush the independents. The opposite happened. Independent restaurants scored 18 points higher on average than national chains.

Chain average: 27/100
Independent average: 46/100
Chains scoring zero: 5 out of 18 (28%)
Independents scoring zero: 6 out of 49 (12%)
Independents with AI bot access: 84%
Chains with AI bot access: 56%

Why are independents winning? Two reasons. First, many independent restaurant sites are built on simpler platforms like Squarespace or WordPress, which produce crawlable HTML by default. Second, chain websites are increasingly built as JavaScript single-page applications that serve empty HTML to crawlers. The fancier the site, the more likely it is to be invisible to AI.

The McDonald's problem

McDonald's, Pizza Hut, Panera Bread, Chick-fil-A, and KFC all scored zero. Their websites are JavaScript applications that return a bare HTML shell with no content to any client that does not execute JavaScript. When GPTBot crawls mcdonalds.com, it receives an empty page. No menu. No locations. No hours. No structured data. Nothing.

This is not a theoretical problem. When a customer asks Perplexity 'What's on the McDonald's menu?' or asks ChatGPT 'Is there a Chick-fil-A near me that's open right now?', these AI systems cannot read the websites. They fall back to whatever third-party data they can find, which is often outdated, incomplete, or wrong. The biggest food brands in the world have made themselves invisible to the fastest-growing search channel.

A JavaScript SPA is not inherently bad for AI search. But it must implement server-side rendering (SSR) or static generation so that crawlers receive fully populated HTML. If your site returns an empty <div id='root'></div> to a bot, you do not exist in AI search. Full stop.

The independent leaderboard

Among the 49 independent restaurants and cafes we audited, a clear pattern emerged. Coffee roasters and established restaurants with proper WordPress or custom builds scored highest. Here are the top performers and what they got right:

Stumptown Coffee (77/100): Full Organization, LocalBusiness, and PostalAddress schema. Sitemap with 6 URLs. All 4 AI bots allowed. llms.txt present.
Portal Tea (73/100): Clean LocalBusiness schema with PostalAddress. Sitemap present. All AI bots allowed.
Puff Coffee (70/100): Organization and WebSite schema with SearchAction. Simple, crawlable HTML.
Jeffrey's Austin (70/100): Full FoodEstablishment schema with PostalAddress and ImageObject. The kind of structured data AI search engines love.
Zen Japanese (69/100): The only site with Restaurant schema that also includes FAQPage, OfferCatalog, and GeoCoordinates. An outlier in the best possible way.

Notice what the top scorers share: structured data that tells AI crawlers exactly what the business is, where it is, and what it offers. They are not doing anything exotic. They have simply completed the basics that 80% of restaurants have not.

The fix: 5 things every restaurant needs

If you run a restaurant, cafe, or catering business, here is your priority list. These are ranked by impact based on what separated high-scoring sites from low-scoring ones in our audit:

1. Add Restaurant or LocalBusiness schema

Only 28% of the sites we audited had any form of LocalBusiness or Restaurant structured data. This is the single most impactful thing you can do. At minimum, add a JSON-LD block with your business name, address, phone, hours, and cuisine type. If you serve food (you are a restaurant, so yes), use the Restaurant schema type, which is a subtype of LocalBusiness and includes fields for menu, price range, and acceptsReservations.

{
  "@context": "https://schema.org",
  "@type": "Restaurant",
  "name": "Your Restaurant Name",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "123 Main St",
    "addressLocality": "Portland",
    "addressRegion": "OR",
    "postalCode": "97201",
    "addressCountry": "US"
  },
  "telephone": "+1-503-555-0100",
  "servesCuisine": ["Italian", "Mediterranean"],
  "priceRange": "$$",
  "openingHours": ["Mo-Sa 11:00-22:00", "Su 12:00-20:00"],
  "acceptsReservations": "True",
  "menu": "https://yourrestaurant.com/menu"
}

2. Add FAQPage schema

This is the highest-leverage structured data type for AI citations and almost nobody has it. Think about the questions customers actually ask: Do you have vegetarian options? Are you open on Mondays? Do you take reservations? Is there parking? Each of these should be a Question/Answer pair in FAQPage schema. When someone asks ChatGPT that same question, your page becomes the citation source.

3. Make sure AI crawlers can read your site

24% of the sites we audited are blocking at least one major AI crawler. Check your robots.txt. If you see Disallow rules for GPTBot, ClaudeBot, PerplexityBot, or Google-Extended, you are telling AI search engines to stay away. Some sites block these bots intentionally, but most do it accidentally through overly aggressive security plugins or template defaults. Every blocked bot is a search engine that cannot cite you.

4. Have a sitemap

Only 64% of audited sites had a sitemap. A sitemap tells crawlers what pages exist. Without one, crawlers have to discover your pages by following links, which means they miss things. If your site is on WordPress, install Yoast SEO or Rank Math. Both generate sitemaps automatically. If you are on Squarespace or Wix, sitemaps are built in. If you have a custom site, generate one at xml-sitemaps.com and submit it to Google Search Console.

5. Create an llms.txt file

33% of independent sites had llms.txt, compared to 11% of chains. llms.txt is a simple text file that tells AI crawlers what your site is about and which pages matter most. It takes 10 minutes to create and it is one of the easiest ways to give AI search engines context about your business. For full instructions, see our llms.txt guide.

What this means for the industry

The restaurant industry is in the middle of a search transition. Customers are shifting from typing queries into Google to asking questions to AI assistants. 'Where should I eat tonight?' is becoming a conversation, not a keyword search. The businesses that win this transition will be the ones that make themselves legible to AI systems. Right now, almost none of them are.

The good news is that the bar is low. A score of 60 out of 100 would put a restaurant in the top tier of the sites we audited. The basics, structured data, crawlable HTML, AI bot access, and FAQPage schema, are not expensive or technically complex. They are simply not being done.

We audited 67 real restaurant websites to produce this report. You can see exactly how your site compares by running a free audit at parceit.com. The audit checks all the same signals: AI crawler access, structured data, sitemap, crawlability, and content structure.

Methodology

All data in this article was collected on July 19, 2026 using the Parceit audit engine. The engine fetches each site's homepage, parses HTML and JSON-LD, checks robots.txt for AI crawler rules, verifies sitemap presence, and scores the site across four categories: Structural (30 points), Entity Recognition (25 points), Content Signals (20 points), and Crawlability (25 points). Chain sites were selected from the top US restaurant chains by revenue. Independent sites were sourced from OpenStreetMap data for Portland OR, Austin TX, and Asheville NC.

Sites scoring zero received empty HTML responses from the server, meaning JavaScript-rendered content that does not server-side render. This is a known limitation of single-page application architectures when no SSR or prerendering fallback is configured.

Frequently asked questions

Why did McDonald's score zero?

McDonald's website is a JavaScript single-page application. When a crawler requests the page, the server returns an empty HTML shell with no content. The actual menu, locations, and hours are rendered by JavaScript in the browser. Since AI crawlers do not execute JavaScript, they see nothing. This is the same issue affecting Pizza Hut, Panera Bread, Chick-fil-A, and KFC.

How can a small restaurant compete with chains on AI search?

Easily. Our data shows independents are already outperforming chains by 18 points on average. A small restaurant with a WordPress or Squarespace site, basic Restaurant schema, and an FAQPage will score higher than McDonald's, Subway, and Chipotle combined. The advantage small businesses have is that their sites are simpler and more crawlable by default.

Is FAQPage really that important?

Yes. When ChatGPT, Perplexity, or Google AI Overviews generates an answer, it looks for authoritative sources to cite. FAQPage schema wraps question-and-answer content in a format the AI can parse with perfect accuracy. Without it, the AI has to infer Q&A from paragraphs of text, which is unreliable. In our audit of 67 sites, only 1 had FAQPage. That site (eatzen.com) is the only restaurant in our sample that would reliably appear as a cited source for question-based queries.

How much does this cost to fix?

For most independent restaurants, the fixes cost nothing if you have a WordPress, Squarespace, or Wix site. Structured data can be added through free plugins. FAQPage schema can be hand-written. llms.txt is a plain text file. The main investment is time, roughly 2 to 4 hours to add schema, write FAQ entries, and verify AI bot access. Parceit offers implementation services starting at 399 dollars if you prefer to have it done professionally.