kennelsonauto.com — AI Site Grade

Analysis

I have enough to write the audit now.

The live site is entirely invisible to AI crawlers — 403 on every page, every bot, every path

The entire domain kennelsonauto.com returns HTTP 403 to every user-agent tested — browsers, GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User, Applebot-Extended, Bytespider, and anthropic-ai — all blocked identically by Cloudflare's challenge wall. The homepage, /robots.txt, /llms.txt, /new-vehicles/, and /used-vehicles/ all serve the same Cloudflare "Attention Required!" HTML shell (4,568 bytes for bots, ~236 KB for browsers due to embedded font data). No AI crawler can access a single byte of real content. There is no robots.txt file to parse — it returns a 403 HTML page, not a text file. There is no llms.txt (also 403). The sitemap is unreachable. The site has effectively zero AI-visibility posture: it is a black box behind Cloudflare's JS-challenge layer.

Cold-Knowledge Gap

A frontier LLM queried cold about "kennelsonauto.com / Kennelson Auto" returned no verifiable information — the model could not confirm services, location, products, or reputation. The brand is entirely absent from the model's training data. This is a complete knowledge vacuum. The gap between what the LLM knows (nothing) and what the site actually represents (a multi-franchise auto group with five dealerships in Dixon, IL) is total. Any AI engine that attempts to retrieve the site for grounding hits a 403 wall and gets nothing.

Content & Schema (Archived Snapshot)

The Wayback Machine snapshot from January 2025 reveals a fully functional site built on the DealerInspire platform (evidenced by the di-uploads-pod47.dealerinspire.com image CDN path). The homepage carried AutoDealer schema with the brand's name ("Ken Nelson Auto Group"), address (1100 N Galena Ave, Dixon, IL 61021), phone (815-288-4455), and price range. The site represented five franchises: Ken Nelson Chevy Buick GMC, Ken Nelson Toyota, Ken Nelson Cadillac, Ken Nelson Nissan, and Ken Nelson CDJR (Chrysler Dodge Jeep Ram). The homepage text positioned the group as serving Sterling, Freeport, and Rochelle with new and used inventory, trade-in tools, financing, and service. The snapshot contained 920 words of visible text, customer reviews from DealerRater, and a clear heading structure. None of this content is accessible on the live site.

External Signals

Web searches for "Kennelson Auto," "Ken Nelson Auto Group," and related queries returned zero indexed results across DuckDuckGo. No reviews, no Reddit threads, no press mentions, no business listings surfaced. The only recoverable evidence of the brand's existence is the Wayback Machine snapshot and the DNS/Cloudflare infrastructure. The site's MX records point to Google Workspace (suggesting active email), and the SPF record references pbssystems.com and email-od.com — likely the dealer management system and email marketing vendors. But none of this is crawlable or citable by AI engines.

The Core Problem

The site is a fully functional auto dealership group with real inventory, five physical locations, and a DealerInspire CMS — but Cloudflare's JS-challenge security layer is blocking all non-human traffic unconditionally. There is no evidence of a deliberate AI-blocking strategy (no robots.txt directives exist because the file itself is blocked). This appears to be a blanket WAF configuration that treats every unrecognized user-agent as a threat. The result: AI crawlers see nothing, LLMs know nothing, and the brand has zero AI-search presence despite having substantial content to offer.

Findings

1. Every page returns HTTP 403 to all user-agents High

   The homepage, /robots.txt, /llms.txt, /new-vehicles/, and /used-vehicles/ all serve a Cloudflare challenge page to every tested bot and browser, blocking all content.

   What to change: Configure Cloudflare WAF to allow AI crawler user-agents (GPTBot, ClaudeBot, etc.) through the JS challenge, or serve static HTML versions to bots.

2. Robots.txt is inaccessible (403) High

   The robots.txt file returns a 403 HTML page instead of a text file, so crawlers cannot read any directives.

   What to change: Ensure robots.txt is served as a plain text file with appropriate allow/disallow rules for AI crawlers.

3. No llms.txt file available Medium

   The llms.txt endpoint returns 403, so AI crawlers cannot discover a curated content summary.

   What to change: Create an llms.txt file listing key pages and content for AI crawlers.

4. LLMs have no knowledge of the brand High

   A frontier LLM queried cold returned no verifiable information about Kennelson Auto, indicating the brand is absent from training data.

   What to change: Allow AI crawlers to index the site so content can appear in future training data.

5. No indexed pages in search engines High

   Web searches for the domain and brand name returned zero results across DuckDuckGo, indicating no pages are indexed.

   What to change: Resolve the 403 issue to allow search engine crawlers to index the site.

6. Sitemap is unreachable Medium

   The sitemap could not be fetched because the entire domain is blocked, preventing crawlers from discovering page structure.

   What to change: Ensure sitemap.xml is accessible and listed in robots.txt.

7. Cloudflare WAF blocks all non-human traffic High

   The site uses Cloudflare's JS-challenge which blocks every unrecognized user-agent, including all AI crawlers, without exception.

   What to change: Create a WAF bypass rule for known AI crawler user-agents or serve static HTML to bots.

8. No schema markup accessible to crawlers High

   Since the live site returns 403, any structured data (like AutoDealer schema) is invisible to AI crawlers.

   What to change: Ensure schema markup is present on live pages and accessible to crawlers.

9. No external citations or reviews indexed Medium

   Web searches found no reviews, press mentions, or business listings for the brand, limiting off-site signals.

   What to change: Encourage customer reviews on third-party sites and ensure they are crawlable.

What's working

• Wayback Machine snapshot shows rich content — A January 2025 snapshot reveals a fully functional site with AutoDealer schema, 920 words of text, and clear structure, proving the site has valuable content.
• Site built on DealerInspire platform — The use of DealerInspire suggests a professional CMS with potential for structured data and SEO features.
• Active email infrastructure via Google Workspace — MX records point to Google Workspace, indicating active business email and potential for communication.