pfaffauto.com — AI Site Grade

Analysis

Pfaffauto.com AI-Visibility Audit

ChatGPT-User, OAI-SearchBot, and PerplexityBot all receive HTTP 429 "Too Many Requests" from a F5 BigIP load balancer when accessing the homepage, while GPTBot, Google-Extended, anthropic-ai, and Perplexity-User pass through with full 200 responses — a selective rate-limit pattern that blocks the very bots most likely to serve AI-generated search answers.

Crawler Access

The robots.txt at pfaffauto.com contains no AI-bot-specific directives whatsoever. No User-agent: GPTBot, no User-agent: Google-Extended, no User-agent: ClaudeBot — the wildcard * rule applies to all, with a 10-second crawl delay and a list of disallowed paths (mostly PDF coupons, CGI scripts, and JSON endpoints). The llms.txt returns a 404 served as a full HTML page (331 KB), meaning the site has no AI-friendly content map. The sitemap index at /sitemaps.xml exists and points to three sub-sitemaps covering roughly 9,300 URLs, including blog posts, inventory pages, and division pages — so Google can discover content, but AI crawlers have no structured guidance.

The compare_bot_access test reveals a split infrastructure: GPTBot, Google-Extended, anthropic-ai, and Perplexity-User all get 200 with full HTML (331 KB). But ChatGPT-User, OAI-SearchBot, and PerplexityBot get 429 from a BigIP server with an 18-byte "Too Many Requests" body. ClaudeBot times out (status 0). This means OpenAI's search-crawler (OAI-SearchBot) and its chat-browsing agent (ChatGPT-User) are blocked, as is Perplexity's primary crawler (PerplexityBot). The BigIP server header on the 429 responses points to an F5 load balancer doing UA-based rate limiting — not a robots.txt block, but equally effective at keeping content out of ChatGPT and Perplexity answers.

Cold-Knowledge Gap

The LLM queried cold knows Pfaff as a "Canadian luxury automotive retailer" representing Porsche, Audi, BMW, McLaren, Lamborghini, and Rolls-Royce, with a "strong motorsport heritage" and 2023 IMSA GTD Pro championship. The actual site reveals a broader and more complex reality that the model misses: Pfaff now operates 15 locations across 12 brands including Honda, Hyundai, Subaru, Volkswagen, Ford, and Harley-Davidson — mainstream and budget brands alongside luxury. The site also discloses a partnership with Lithia and Driveway (NYSE: LAD), North America's second-largest publicly traded dealership group — a structural fact absent from the model's prior. The cold knowledge also omits the Pfaff Concierge, Pfaff Autoworks, Pfaff Leasing, Pfaff Reserve, Pfaff Tuning, and Pfaff Fleet divisions, which are prominently listed on the site.

Schema Posture

Every page carries the same AutoDealer + AutomotiveBusiness + WebSite JSON-LD block with identical geo-coordinates, address, phone, and social links. The WebSite schema includes a SearchAction targeting the used-inventory search endpoint. However, no page-level schema exists for individual vehicle inventory pages, blog articles, or FAQ content. The blog posts (e.g., the 2022 IMSA championship article) carry only BreadcrumbList plus the generic dealer schema — no Article or NewsArticle type. The homepage has no FAQ schema despite the site using comparison language and list structures that would benefit from it. The priceRange field is set to "$-

Content & Technical Signals

The homepage renders 420 words of visible text from a plain GET — no JS-rendering dependency, which is good for crawlers. But the meta description on the homepage and the 404 page both contain empty brand-name placeholders (" dealer in Toronto. Come see our rebates..."), suggesting a template rendering issue where brand names fail to populate. The og:type is set to "product" on every page including the homepage, which is semantically incorrect for a dealership group. The site runs on a D2C Media platform (evident from CDN domains, footer credit, and resource loader patterns) with F5 BigIP load balancing and Google Tag Manager. The viewport is hardcoded to width=960 — a desktop-only constraint that may affect mobile usability signals.

Findings

1. ChatGPT-User, OAI-SearchBot, and PerplexityBot blocked by HTTP 429 High

   The F5 BigIP load balancer returns HTTP 429 'Too Many Requests' for ChatGPT-User, OAI-SearchBot, and PerplexityBot, while GPTBot, Google-Extended, anthropic-ai, and Perplexity-User receive full 200 responses. This selective rate-limiting blocks the very bots most likely to serve AI-generated search answers, preventing Pfaffauto.com content from appearing in ChatGPT and Perplexity responses.

   What to change: Remove the F5 BigIP rate-limit rules that target ChatGPT-User, OAI-SearchBot, and PerplexityBot, or whitelist these bots to allow full access.

2. No AI-bot-specific directives in robots.txt Medium

   The robots.txt file contains no User-agent lines for GPTBot, Google-Extended, ClaudeBot, or any other AI crawler. The wildcard rule applies to all bots with a 10-second crawl delay and disallowed paths, but no explicit guidance exists for AI crawlers, leaving them to rely on the generic rules.

   What to change: Add explicit User-agent directives for GPTBot, Google-Extended, ClaudeBot, and other AI crawlers to control their access and crawl rate.

3. llms.txt returns 404 Medium

   The llms.txt file at pfaffauto.com/llms.txt returns a 404 status, served as a full HTML page (331 KB). This means the site provides no AI-friendly content map or structured guidance for LLMs to discover and understand its content.

   What to change: Create an llms.txt file that lists key pages (e.g., brands, divisions, inventory) and provides a brief summary of the site's content for AI crawlers.

4. ClaudeBot times out on homepage High

   When fetching the homepage as ClaudeBot, the request times out (ReadTimeout error), indicating that ClaudeBot cannot access the site at all. This prevents Claude from indexing Pfaffauto.com content.

   What to change: Investigate and resolve the timeout issue for ClaudeBot, possibly by adjusting server or firewall settings to allow ClaudeBot's requests.

5. Cold LLM knowledge omits mainstream brands and Lithia partnership Medium

   The LLM's cold knowledge describes Pfaff as a luxury retailer (Porsche, Audi, BMW, etc.) but misses that the site now represents 12 brands including Honda, Hyundai, Subaru, Volkswagen, Ford, and Harley-Davidson. The partnership with Lithia and Driveway (NYSE: LAD) is also absent. This gap means AI-generated answers about Pfaff will be incomplete.

   What to change: Ensure that all brand pages and the about page clearly list all brands and the Lithia partnership in structured data and prominent text to improve LLM knowledge extraction.

6. Cold LLM knowledge omits key divisions Medium

   The LLM's cold knowledge does not mention Pfaff Concierge, Pfaff Autoworks, Pfaff Leasing, Pfaff Reserve, Pfaff Tuning, or Pfaff Fleet, even though these divisions are prominently listed on the site. This limits the depth of AI-generated answers about Pfaff's services.

   What to change: Add structured data (e.g., Organization schema with department sub-properties) for each division on the divisions page and link to them from the homepage.

7. No page-level schema for inventory, articles, or FAQ High

   Every page carries the same generic AutoDealer + AutomotiveBusiness + WebSite JSON-LD block. Individual vehicle inventory pages, blog articles, and FAQ content lack specific schema types (e.g., Product, Article, FAQPage). This reduces the chances of rich results in AI-generated answers.

   What to change: Add Product schema to inventory pages, Article schema to blog posts, and FAQPage schema to any FAQ content. Use unique schema per page type.

8. priceRange field set to placeholder value Low

   The JSON-LD schema includes a priceRange field set to "$-

   quot;, which conveys no actual pricing signal. This placeholder value does not help AI systems understand the dealership's pricing tiers.

   What to change: Update the priceRange field to reflect actual pricing, e.g., "$20,000 - $200,000" or similar range based on inventory.

9. Meta descriptions contain empty brand-name placeholders Medium

   The homepage and 404 page meta descriptions contain empty placeholders like " dealer in Toronto. Come see our rebates...", indicating a template rendering issue where brand names fail to populate. This results in broken snippets in search results.

   What to change: Fix the template rendering logic to ensure brand names are correctly inserted into meta descriptions on all pages.

10. og:type set to 'product' on homepage Low

    The Open Graph type is set to 'product' on every page, including the homepage, which is semantically incorrect for a dealership group. This may confuse social media crawlers and AI systems that parse Open Graph data.

    What to change: Change the og:type on the homepage to 'website' and on other pages to appropriate types (e.g., 'article' for blog posts).

11. Viewport hardcoded to width=960 Low

    The viewport meta tag is set to 'width=960', a desktop-only constraint that may negatively affect mobile usability signals and mobile-first indexing.

    What to change: Change the viewport meta tag to 'width=device-width, initial-scale=1' to support responsive design.

12. No web search results for Pfaffauto.com or related queries High

    Web searches for 'pfaffauto.com', 'Pfaff Automotive Partners reviews reputation Canada', and 'Pfaff Automotive Lithia Driveway partnership' returned zero results. This indicates very low external visibility and backlink profile, limiting the site's authority for AI systems.

    What to change: Investigate and improve the site's backlink profile and online presence through PR, partnerships, and content marketing to increase external signals.

What's working

• GPTBot and Google-Extended allowed full access — GPTBot and Google-Extended receive 200 responses with full HTML content, meaning OpenAI's training crawler and Google's AI crawler can index the site without restrictions.
• Sitemap index with 9,300 URLs available — The sitemap index at /sitemaps.xml points to three sub-sitemaps covering approximately 9,300 URLs, including blog posts, inventory pages, and division pages, enabling good discoverability for crawlers that can access it.
• Homepage renders 420 words of visible text without JavaScript — The homepage returns 420 words of visible text from a plain GET request, with no JavaScript rendering dependency, ensuring that all crawlers can read the content.
• JSON-LD schema with AutoDealer and SearchAction present — Every page includes a JSON-LD block with AutoDealer, AutomotiveBusiness, and WebSite schema, including a SearchAction targeting the used-inventory search endpoint, providing basic structured data.
• Brands and divisions pages contain detailed text content — The 'Our Brands' and 'Our Divisions' pages contain over 1,300 words each, listing all brands and divisions with descriptions, providing rich content for AI crawlers to extract.
• Motorsport championship article published on site — The site hosts a detailed article about Pfaff Motorsports winning the 2022 IMSA GTD Pro Championship, providing authoritative content that can be indexed and cited.
• About page provides comprehensive company overview — The about-us page contains nearly 2,000 words covering the company's history, values, and the Lithia partnership, offering substantial context for AI systems.
• Why Buy From Pfaff page provides value proposition — The 'Why Buy From Pfaff' page contains 1,195 words detailing the dealership's advantages, which can help AI systems understand the brand's unique selling points.