nelsoncars.com — AI Site Grade

Analysis

The llms.txt URL silently redirects to the homepage — no dedicated file exists, meaning every AI crawler that requests it gets the full homepage HTML instead of a structured content map.

Crawler Access

The robots.txt is unusually comprehensive for a regional dealership group. It explicitly names GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, Perplexity-User, and Google-Extended — all with Allow: /. Every AI bot tested (11 UAs) returns a 200 with full HTML content (~215KB), identical to browser baseline. No UA-based blocking, no Cloudflare challenge, no JS shell. The site runs on Fastly CDN (Kestrel server, Varnish cache) behind Google infrastructure. The /llms.txt endpoint does not exist; it 200-redirects to the homepage, which is a missed opportunity for structured AI content discovery.

Cold-Knowledge Gap

A frontier LLM queried cold about "Nelson Cars automotive group" has no specific, verified information — it describes the brand as "not a widely recognized major automotive manufacturer or retailer" and suggests it "may be a small, regional dealership group." This is the critical gap: the site itself describes a 50+ year history (founded in the 1950s as Jim Nelson Ford), a 6-location group across Oklahoma, Tennessee, and Arkansas, with Nissan and Mazda franchises, a used-car arm (Car Country), and a branded sales philosophy ("The Nelson Difference"). None of this substance exists in the model's prior knowledge.

Schema Posture

Every page carries an AutoDealer JSON-LD block, but it is identical and incomplete across all pages — same Tulsa address, same empty telephone fields ("+1 "), same empty sameAs array, same openingHoursSpecification: []. The inventory pages add rich Vehicle schema with VIN, price, mileage, and condition — these are well-structured. But the core AutoDealer schema lacks phone numbers, social profiles, and hours, which are all present in the visible HTML on the /service and /finance pages. The priceRange of "$1200-$15000" is misleadingly low for a group selling new 2026 Mazda CX-90s at $37,360+.

Content Architecture

The site has 9 content pages (about, contact, service, finance, the-nelson-difference, the-nelson-story, ministry-partners, mazda-locations, shop-from-home) plus 1,850+ inventory URLs across new, used, and CPO. The "The Nelson Story" page is the strongest AI-facing asset — it contains a detailed 50-year timeline, founder biography, and the "One Price, One Person" philosophy. The "Ministry Partners" page is unusual for a dealership: it lists 10 Christian nonprofit partners with external links, signaling a faith-driven brand identity that no AI model currently knows. No FAQ schema, no comparison tables, no structured Q&A content exists anywhere on the site.

External Signals

Web search returned zero indexed results for "Nelson Cars automotive group reviews Tulsa Oklahoma" — no Google Business Profile snippets, no Reddit threads, no press mentions surfaced. The site has a Google Search Console verification TXT record (google-site-verification=8xrwqyy-oHFySk6dmI4dNELLQyt4ml754ofpP-Wkm2w) but no other external citation footprint. The individual dealership domains (nelsonnissan.com, nelsonmazdaok.com, etc.) are separate sites that handle actual transactions — the nelsoncars.com domain functions as a thin corporate umbrella with no blog, no news, no reviews, and no user-generated content for AI engines to cite.

Findings

1. Frontier LLMs have no knowledge of Nelson Cars High

   A cold query about 'Nelson Cars automotive group' returned no specific information; the model described it as a small, regional dealership group. The site's 50+ year history, six locations, and brand philosophy are absent from AI knowledge.

   What to change: Publish a comprehensive llms.txt file and ensure key pages (history, locations, values) are crawlable and well-structured with schema markup to seed AI knowledge.

2. llms.txt endpoint redirects to homepage High

   The /llms.txt URL returns a 200 redirect to the homepage instead of a dedicated file, so AI crawlers receive full homepage HTML rather than a structured content map.

   What to change: Create a valid llms.txt file at /llms.txt that lists key pages and their descriptions for AI crawlers.

3. AutoDealer schema is identical and incomplete across all pages High

   Every page carries an AutoDealer JSON-LD block with the same Tulsa address, empty telephone fields, empty sameAs array, and empty openingHoursSpecification. Phone numbers and hours exist in visible HTML but are missing from schema.

   What to change: Populate telephone, sameAs, and openingHoursSpecification fields in the AutoDealer schema with actual data from the site.

4. AutoDealer priceRange is misleadingly low Medium

   The priceRange field is set to '$1200-$15000', but the site sells new 2026 Mazda CX-90s at $37,360+. This misrepresents the dealership's pricing to AI systems.

   What to change: Update the priceRange to reflect actual inventory pricing, e.g., '$1200-$50000' or remove it if dynamic.

5. No FAQ or Q&A schema on any page Medium

   The site lacks FAQ schema, comparison tables, or structured Q&A content, missing opportunities to appear in AI-generated answers and rich snippets.

   What to change: Add FAQ schema to pages like finance, service, and shop-from-home to answer common customer questions.

6. No external citations or reviews indexed High

   Web search returned zero indexed results for 'Nelson Cars automotive group reviews Tulsa Oklahoma'. No Google Business Profile snippets, Reddit threads, or press mentions were found.

   What to change: Encourage customer reviews on Google and other platforms, and publish press releases or blog posts to build external citation footprint.

7. nelsoncars.com is a thin corporate umbrella with no blog or news Medium

   The domain functions as a corporate umbrella with only 9 content pages and no blog, news, or user-generated content for AI engines to cite. Individual dealerships operate on separate domains.

   What to change: Add a blog or news section to nelsoncars.com to publish updates, community involvement, and dealership news.

8. Inventory sitemap returns empty content Medium

   The /inventorysitemap.xml and /pagelinkssitemap.xml URLs return 200 but with 0 words, providing no URLs to crawlers.

   What to change: Populate inventorysitemap.xml and pagelinkssitemap.xml with actual URLs to ensure all inventory pages are discoverable.

What's working

• robots.txt explicitly allows all major AI bots — The robots.txt names 8 AI bots (GPTBot, OAI-SearchBot, etc.) with Allow: /, ensuring full crawl access for AI crawlers.
• AI bots receive full HTML content, not JS shells — All 11 tested AI bots receive a 200 response with full HTML (~215KB), identical to browser baseline, with no UA-based blocking or JavaScript rendering issues.
• Inventory pages include rich Vehicle schema — New and used inventory pages contain Vehicle JSON-LD with VIN, price, mileage, and condition, providing structured data for AI systems.
• The Nelson Story page provides rich historical content — The 719-word page includes a 50-year timeline, founder biography, and brand philosophy, offering substantial material for AI knowledge extraction.
• Ministry Partners page signals faith-driven brand identity — The page lists 10 Christian nonprofit partners with external links, providing unique brand context that AI models currently lack.
• Site is verified in Google Search Console — A TXT record confirms Google Search Console verification, indicating the site is actively monitored for search performance.