swickard.com — AI Site Grade

Analysis

Swickard.com — AI-Visibility Audit

The site's corporate homepage (swickard.com) is blocked to GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Bytespider, and anthropic-ai — all returning 0-byte errors — while the actual inventory subdomain (swickardautogroup.com) serves all bots full content with rich AutomotiveBusiness schema. This creates a fragmented AI footprint where the brand story is invisible to most crawlers but the product catalog is fully exposed.

Crawler Access

The robots.txt at swickard.com is a bare Yoast-generated file with a single User-agent: * Disallow: rule and no AI-bot directives whatsoever. Despite the permissive robots.txt, the nginx server behind swickard.com silently drops connections from GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Bytespider, and anthropic-ai (status 0, 0 bytes). ChatGPT-User, OAI-SearchBot, PerplexityBot, and Perplexity-User all get the full 138KB page. The inventory subdomain at www.swickardautogroup.com — a separate DealerOn-hosted platform — serves all bots identically with 527KB of content and AutomotiveBusiness + ItemList schema. No llms.txt exists (404). The site runs on nginx with no visible WAF headers, hosted at a Google Cloud IP (35.215.74.230).

Cold-Knowledge Gap

The LLM's prior knowledge describes Swickard as a "privately held dealership network" with "30+ dealerships by 2023" that faced a "class-action lawsuit alleging deceptive pricing." The actual site claims 50+ dealerships across six states (including Texas, which is not in the cold knowledge) and promotes a Swickard Price Promise — a 3-day price-match guarantee — that directly contradicts the lawsuit narrative. The site makes zero mention of any lawsuit, hidden fees, or no-haggle pricing. The cold model knows about the lawsuit; the site is silent on it. The model also undercounts the dealership count by nearly 40%.

Content & Schema Posture

The homepage and all corporate pages use only WebPage, WebSite, Organization, and BreadcrumbList schema — no AutomotiveBusiness, AutoDealer, FAQPage, Product, or LocalBusiness types. The inventory subdomain at swickardautogroup.com uses AutomotiveBusiness and ItemList schema with rich vehicle data, but this is a completely separate domain with no schema bridging back to the corporate brand. The /state/texas/ page is a blog archive with no Texas dealership listings and is marked noindex. The /about-us/ page returns only 47 words of visible text — essentially a JavaScript-rendered shell with minimal content for crawlers. The /who-is-jeff-swickard/ article has a proper Article schema with 1,281 word count and an FAQ section, making it the single most AI-friendly page on the domain.

External Signals

The press page lists 20+ real press mentions from CBT News, Portland Business Journal, PR Newswire, and industry podcasts — including a 2026 acquisition of a downtown Portland high-rise and a Reynolds & Reynolds AI partnership. No negative press, lawsuit references, or Reddit threads surfaced in search. The cold model's lawsuit knowledge appears stale or from a source no longer indexed. The DNS includes an anthropic-domain-verification TXT record, confirming the brand has explicitly verified ownership with Anthropic — yet ClaudeBot still gets blocked at the server level.

Findings

1. Server silently drops connections from six major AI crawlers High

   The nginx server behind swickard.com returns 0-byte errors for GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Bytespider, and anthropic-ai, despite a permissive robots.txt. ChatGPT-User, OAI-SearchBot, PerplexityBot, and Perplexity-User are allowed.

   What to change: Remove server-level blocking for AI crawlers or configure the server to serve content to all legitimate bots.

2. LLM prior knowledge contradicts current site claims High

   The cold model describes Swickard as having 30+ dealerships by 2023 and a class-action lawsuit over deceptive pricing. The site claims 50+ dealerships and promotes a price-match guarantee, with no mention of any lawsuit. The model undercounts dealerships by nearly 40%.

   What to change: Ensure the site prominently and accurately represents current dealership count and addresses any negative narratives to correct AI knowledge.

3. Corporate site lacks AutomotiveBusiness and AutoDealer schema Medium

   The homepage and corporate pages use only WebPage, WebSite, Organization, and BreadcrumbList schema. No AutomotiveBusiness, AutoDealer, FAQPage, or Product schema is present, limiting AI understanding of the business type.

   What to change: Add AutomotiveBusiness or AutoDealer schema to the homepage and relevant pages.

4. About Us page has only 47 words of visible text Medium

   The /about-us/ page returns only 47 words of visible text, appearing as a JavaScript-rendered shell with minimal content for crawlers.

   What to change: Expand the About Us page with substantial text content about the company history, values, and leadership.

5. No llms.txt file available for AI crawlers Low

   The site returns a 404 for /llms.txt, missing an opportunity to guide AI crawlers to key pages and context.

   What to change: Create an llms.txt file listing important pages and a brief overview of the company.

6. Texas state page is marked noindex with no dealership listings Medium

   The /state/texas/ page is a blog archive with no Texas dealership listings and is marked noindex, making it invisible to search engines and AI crawlers.

   What to change: Remove the noindex directive and add actual dealership listings or location pages for Texas.

7. Inventory subdomain is a separate domain with no schema bridging Medium

   The inventory subdomain (www.swickardautogroup.com) uses AutomotiveBusiness and ItemList schema but is a completely separate domain with no schema linking back to the corporate brand, fragmenting the AI footprint.

   What to change: Add sameAs or isPartOf schema properties to bridge the inventory subdomain to the corporate domain.

8. ClaudeBot blocked despite anthropic-domain-verification TXT record High

   The DNS includes an anthropic-domain-verification TXT record, confirming the brand has explicitly verified ownership with Anthropic, yet ClaudeBot still gets blocked at the server level.

   What to change: Allow ClaudeBot access at the server level to honor the domain verification.

9. Blog page contains only 16 words of visible text Medium

   The /blog/ page returns only 16 words of visible text, likely a JavaScript shell with minimal content for crawlers.

   What to change: Ensure the blog page renders full content server-side or provides a static fallback for crawlers.

10. Price Promise page lacks FAQ schema for key questions Low

    The Swickard Price Promise page (525 words) explains a 3-day price-match guarantee but uses no FAQPage schema, missing an opportunity to surface common questions in AI answers.

    What to change: Add FAQPage schema to the Price Promise page with common questions and answers.

What's working

• Inventory subdomain serves all bots full content with rich schema — The www.swickardautogroup.com subdomain returns 527KB of content to all 11 tested bots and includes AutomotiveBusiness and ItemList schema with rich vehicle data.
• Who Is Jeff Swickard article has Article schema and FAQ section — The /who-is-jeff-swickard/ page has proper Article schema with 1,281 words and an FAQ section, making it the most AI-friendly page on the domain.
• Press page lists 20+ real press mentions from reputable sources — The press page includes mentions from CBT News, Portland Business Journal, PR Newswire, and industry podcasts, providing strong external signals.
• Domain has anthropic-domain-verification TXT record — The DNS includes an anthropic-domain-verification TXT record, confirming the brand has explicitly verified ownership with Anthropic.
• Robots.txt allows all user agents with no disallowed paths — The robots.txt file is permissive with a single User-agent: * Disallow: rule, allowing all bots to crawl the entire site.
• Heart of Swickard page has substantial text content — The /heart-of-swickard/ page contains 508 words of text, providing meaningful content about company culture.
• California dealerships page has 345 words of content — The /state/california/ page contains 345 words and lists dealerships, providing useful location content.
• No negative press or lawsuit references found in web search — Web searches for lawsuits, deceptive pricing, and Reddit discussions returned zero results, indicating no current negative external signals.