vanceautogroup.com — AI Site Grade

Analysis

I have enough data now. Let me compile the audit.

Cold-Knowledge Gap: LLM thinks Vance Auto Group is in Pennsylvania/Ohio selling GM vehicles — the site is in Oklahoma selling Ford, Ram, Jeep, and Chevy

The AI model's prior knowledge about Vance Auto Group is entirely wrong on geography and brand mix. The cold LLM described it as a "network of car dealerships primarily located in Pennsylvania and Ohio, selling new and used vehicles from brands like Chevrolet, Buick, GMC, and Cadillac." The actual site operates 11 dealerships across Guthrie, Miami, and Woodward, Oklahoma, selling Ford, Lincoln, Chevy, Buick, GMC, Chrysler, Dodge, Jeep, Ram, and Wagoneer. This is a severe knowledge gap — any AI engine answering a query about this group without live retrieval will mislocate it by over 1,000 miles.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, Bytespider — receive a 200 with full HTML content on the www subdomain. No UA-based blocking exists. The robots.txt uses a single User-Agent: * rule with no AI-specific directives; it blocks only internal AJAX endpoints and RSS feeds. The bare domain vanceautogroup.com has an SSL certificate mismatch (cert is for www.vanceautogroup.com), causing all HTTPS requests to fail. HTTP redirects to the www version, so crawlers that follow redirects will reach content — but the bare domain itself is a dead end for any bot that strictly validates TLS.

Content & Schema Posture

The homepage carries AutomotiveBusiness, WebSite, and Product (with AggregateRating: 4.9 / 2,587 reviews) JSON-LD schema. However, the AutomotiveBusiness description contains a template placeholder — "We specialize in new Group and used %(VEHICLE_MAKE), service, and financing." — an unpopulated variable that renders as literal text in the structured data. The site runs on DealerOn (ASP.NET + nginx + Varnish CDN). No llms.txt exists (404). The sitemap timed out on fetch. Multiple navigation-linked pages return 404s: /about-us.html, /our-group, /community-service, /careers — these are listed in the global navigation but have no content. The homepage claims "9 full-service dealerships" while the meta description says "11 dealerships" — an internal inconsistency.

External Signals

The site has near-zero indexed external footprint. Searches for "Vance Auto Group Oklahoma," "John Vance Auto Group reviews," and site:vanceautogroup.com returned zero results from DuckDuckGo. The only external references are the social media links in the schema (Facebook, Twitter, Instagram, YouTube, Cars.com) and a sponsorship link to madepossibleby.us. No Reddit threads, news articles, or third-party review aggregator pages surfaced. The group's fleet and mobility divisions live on separate domains (vancefleet.com, newbyvancemobility.com) with no cross-linking from the main site's content body.

Surprising Findings

The homepage copyright reads "Copyright (c) 2026" — two years in the future — suggesting the DealerOn platform auto-increments the year without editorial review. The Product schema's aggregateRating (4.9 / 2,587 reviews) is a strong trust signal but appears identically on every page including 404s, which may dilute its credibility with crawlers that detect context mismatch. The site promotes a Lifetime Powertrain Protection program as a key differentiator, yet the terms page lists extensive exclusions (luxury brands, commercial vehicles, modified vehicles, branded titles) that could surprise customers — and this nuance is absent from the homepage's structured data.

Findings

1. LLM prior knowledge mislocates Vance Auto Group by over 1,000 miles High

   AI models believe Vance Auto Group is in Pennsylvania/Ohio selling GM vehicles, but the actual site operates 11 dealerships in Oklahoma selling Ford, Ram, Jeep, and Chevy. This mismatch causes AI answers to be factually wrong without live retrieval.

   What to change: Publish a structured knowledge profile (llms.txt, Wikipedia, or schema.org/WebSite with location and brand details) to correct the AI model's prior.

2. AutomotiveBusiness schema contains unpopulated template variable High

   The JSON-LD description reads 'We specialize in new Group and used %(VEHICLE_MAKE), service, and financing.' — the placeholder %(VEHICLE_MAKE) is rendered literally, making the structured data nonsensical.

   What to change: Replace the template variable with actual vehicle makes (e.g., Ford, Ram, Jeep, Chevy) in the schema description.

3. Multiple navigation-linked pages return 404 errors High

   Pages for /about-us.html, /our-group, /community-service, /careers are listed in the global navigation but return 404s, creating dead ends for users and crawlers.

   What to change: Restore these pages with actual content or remove them from the navigation to avoid broken links.

4. Site has almost no indexed external presence High

   Searches for the brand name, reviews, and site: queries returned zero results on DuckDuckGo. No third-party articles, forum mentions, or review pages were found, limiting AI citation sources.

   What to change: Build external citations through local business listings, press releases, and customer review platforms to increase indexed references.

5. Bare domain SSL certificate mismatch blocks HTTPS access Medium

   The certificate for vanceautogroup.com is issued to www.vanceautogroup.com, causing all HTTPS requests to the bare domain to fail with SSL errors. HTTP redirects to www, but strict TLS-validating bots may not follow.

   What to change: Obtain a certificate that covers both the bare domain and www subdomain, or configure a redirect that works over HTTPS.

6. No llms.txt file published for AI guidance Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a curated list of important pages and context.

   What to change: Create an llms.txt file listing key pages (inventory, about, contact) and a brief summary of the business.

7. Sitemap fetch times out, hindering discovery Medium

   The sitemap at /sitemap.xml timed out on multiple attempts, preventing crawlers from efficiently discovering all pages.

   What to change: Ensure the sitemap is accessible and returns quickly; consider generating a static XML file.

8. Homepage claims 9 dealerships, meta description says 11 Low

   The homepage text states '9 full-service dealerships' while the meta description says '11 dealerships', creating confusion for users and AI.

   What to change: Reconcile the count to a single accurate number across all site content.

9. Copyright year shows 2026, two years in the future Low

   The homepage footer reads 'Copyright (c) 2026', which may appear as an error to users and crawlers.

   What to change: Update the copyright year to the current year or use a dynamic range (e.g., 2024-2025).

10. Aggregate rating schema appears on 404 pages, diluting trust Low

    The Product schema with aggregateRating (4.9 / 2,587 reviews) is served on 404 error pages, which may confuse crawlers and reduce credibility.

    What to change: Remove structured data from error pages or ensure it only appears on relevant content pages.

What's working

• All major AI crawlers receive full HTML content — GPTBot, ClaudeBot, PerplexityBot, and others get a 200 response with complete HTML on the www subdomain, with no UA-based blocking.
• robots.txt does not block any AI bots — The robots.txt uses a single User-Agent: * rule with no AI-specific directives, allowing all crawlers to access the site.
• Homepage includes AutomotiveBusiness, WebSite, and Product schema — The homepage has structured data marking it as an AutomotiveBusiness with aggregate rating, which helps AI understand the business type and reputation.
• Product schema shows 4.9 rating from 2,587 reviews — The aggregateRating in schema is a strong trust signal, indicating high customer satisfaction.
• Schema includes links to Facebook, Twitter, Instagram, YouTube, Cars.com — These external profile links provide additional citation sources for AI models.
• Dedicated page for Lifetime Powertrain Protection program — The site has a detailed page explaining the program, which can be a unique selling point for AI-generated answers.