gunnauto.com — AI Site Grade

Analysis

Gunn Automotive Group: A Site That Only Exists for Three AI Crawlers

The homepage of gunnauto.com returns a 403 Cloudflare block to every browser, every search engine crawler, and every major AI bot except three: ChatGPT-User, OAI-SearchBot, and Perplexity-User all receive a 200 with full HTML content. Every other bot — GPTBot, ClaudeBot, Google-Extended, PerplexityBot, Applebot-Extended, Bytespider — gets the same Cloudflare block page as a human visitor.

Crawler Access

The site runs on Cloudflare with a WAF rule that appears to whitelist only OpenAI's and Perplexity's user-agent strings. robots.txt is unreachable (403), llms.txt is unreachable (403), and the XML sitemap at /sitemap.xml is also 403'd. The DNS resolves to Cloudflare IPs (104.17.37.150, 104.17.38.150), and the site is built on the Dealer eProcess platform (cdn.dealereprocess.org). The three allowed bots see a fully rendered HTML page with inventory listings, but the site is effectively invisible to Google's crawler (Google-Extended blocked) and to every other major AI training crawler.

Cold-Knowledge Gap

The LLM queried cold about gunnauto.com returned zero knowledge — no awareness of the brand, its products, its locations, or its reputation. The site presents itself as Gunn Automotive Group, a multi-franchise dealer group in San Antonio, TX selling Acura, Buick, Chevrolet, Chrysler, Dodge, GMC, Honda, Jeep, Nissan, and Ram. The gap between the site's self-description (a 10-brand dealer group with sales, service, finance, and parts) and the LLM's total ignorance is absolute. The brand has no detectable external web presence: searches for "Gunn Automotive" across multiple queries returned zero results from any index.

Schema Posture

The homepage and all subpages contain zero JSON-LD structured data. No AutoDealer, Organization, LocalBusiness, Product, or Vehicle schema is present in any fetched page. The only metadata is basic HTML <meta> tags with keywords and descriptions. The inventory search pages carry <meta name="robots" content="noindex"> and canonical URLs pointing to a generic /search/?cy=78205 — meaning the vehicle inventory, the core commercial content, is explicitly excluded from search indexes.

Content Structure and Broken Pages

The site has a handful of functional pages (homepage, about-us, contact-us, service-departments, finance-departments, parts-departments, locations-and-directions, privacy-policy, sitemap) but multiple critical pages return 404: /inventory, /service, /parts, /finance, /honda, /chevrolet. The site's navigation likely relies on JavaScript to load content, but the HTML served to bots is a heavy JS shell (150-270KB per page) with minimal visible text. No FAQ, comparison tables, or answer-format signals were detected on any page. The contact page lists a sales phone (888-452-2856), service phone (210-239-8914), and parts phone (210-405-5808), but no physical street address is visible in the fetched HTML excerpts.

Findings

1. Cloudflare WAF blocks all major AI crawlers except three High

   The site returns a 403 Cloudflare block to GPTBot, ClaudeBot, Google-Extended, PerplexityBot, Applebot-Extended, and Bytespider. Only ChatGPT-User, OAI-SearchBot, and Perplexity-User receive a 200 with full HTML content.

   What to change: Remove the Cloudflare WAF rule that blocks AI crawlers, or replace it with a rate-limiting approach that allows all legitimate bots.

2. robots.txt returns 403 and is inaccessible High

   The robots.txt file at /robots.txt returns a 403 Cloudflare block, preventing crawlers from reading any directives.

   What to change: Ensure robots.txt is publicly accessible and includes directives that allow AI crawlers.

3. XML sitemap returns 403 and is inaccessible High

   The sitemap at /sitemap.xml returns a 403 Cloudflare block, preventing crawlers from discovering the site's URL structure.

   What to change: Make the sitemap publicly accessible and submit it to search engines.

4. llms.txt returns 403 and is inaccessible Medium

   The llms.txt file at /llms.txt returns a 403 Cloudflare block, preventing AI crawlers from discovering the site's AI-friendly content.

   What to change: Create a publicly accessible llms.txt file that lists the site's key pages and structured data endpoints.

5. No JSON-LD structured data on any page High

   The homepage and all subpages contain zero JSON-LD structured data. No AutoDealer, Organization, LocalBusiness, Product, or Vehicle schema is present.

   What to change: Add JSON-LD structured data for AutoDealer, Organization, LocalBusiness, and Vehicle/Product schemas on all relevant pages.

6. Inventory search pages have noindex and generic canonical URLs High

   Inventory search pages carry <meta name="robots" content="noindex"> and canonical URLs pointing to a generic /search/?cy=78205, excluding vehicle inventory from search indexes.

   What to change: Remove the noindex directive from inventory pages and use unique, descriptive canonical URLs for each search result page.

7. Multiple critical pages return 404 errors High

   Pages for /inventory, /service, /parts, /finance, /honda, /chevrolet, /hours, and /service-center all return 404 errors, indicating broken navigation or missing content.

   What to change: Fix the routing or create proper pages for these URLs, or redirect them to the correct functional pages.

8. Brand has no detectable external web presence High

   Searches for 'Gunn Automotive' across multiple queries returned zero results from any search index, and the LLM had no prior knowledge of the brand.

   What to change: Build external citations through Google Business Profile, industry directories, and local SEO efforts to establish the brand's online presence.

9. No physical street address visible in fetched HTML Medium

   The contact page lists phone numbers but no physical street address is visible in the HTML excerpts, which is critical for LocalBusiness schema and local SEO.

   What to change: Add the full physical address to the contact page and include it in structured data.

10. Pages are heavy JS shells with minimal visible text Medium

    Fetched pages are 150-270KB with minimal visible text, suggesting heavy reliance on JavaScript for content rendering, which may not be fully executed by all crawlers.

    What to change: Ensure critical content is server-side rendered or included in the initial HTML payload to be accessible to all crawlers.

11. No FAQ, comparison tables, or answer-format signals detected Medium

    The site lacks FAQ sections, comparison tables, or other answer-format content that AI crawlers often use to extract structured answers.

    What to change: Add FAQ schema and comparison content for vehicle models and services to improve AI visibility.

What's working

• Site allows ChatGPT-User, OAI-SearchBot, and Perplexity-User crawlers — The site returns full HTML content to ChatGPT-User, OAI-SearchBot, and Perplexity-User, enabling these AI assistants to index the site's content.
• Key pages like about-us, contact-us, and service departments are accessible — Pages for about-us, contact-us, service-departments, finance-departments, parts-departments, locations-and-directions, and privacy-policy return 200 and contain content.
• Inventory search pages render full vehicle listings to allowed bots — The new and used inventory search pages return large HTML payloads (2.7MB) with vehicle listings, providing rich content to allowed AI crawlers.
• HTML sitemap page is accessible and contains links — The /sitemap page returns a 200 with 579KB of content, providing a human-readable sitemap that may help crawlers discover pages.
• Branded search pages for Honda and Acura are accessible — Pages like /search/new-honda-san-antonio-tx and /search/new-acura-san-antonio-tx return 200 with inventory content, targeting specific makes.
• Subdomains for Honda and Acura resolve to A records — Subdomains honda.gunnauto.com and acura.gunnauto.com have DNS A records, indicating potential for brand-specific microsites.