shopkunes.com — AI Site Grade

Analysis

The Midwest's Best Dealership Group has zero cold LLM awareness despite operating 40+ locations and selling 23,000+ vehicles annually — a complete AI-visibility blind spot for a major regional auto group.

Crawler Access

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, Bytespider — receives a full 200 response with identical byte payload (~1.9 MB) to a browser baseline. No UA-based blocking exists. The site runs on Google Frontend (GCP) via a Next.js SSR stack (RideMotive platform), so server-rendered HTML is served to all bots. The robots.txt contains no AI-bot-specific directives; only dotbot and MotoMinerBot are blocked entirely. An llms.txt exists and is massive (270 KB), listing thousands of inventory URLs — but every entry uses the same generic description "New quality vehicles in Delavan," making it a low-signal directory dump rather than a curated AI resource.

Cold-Knowledge Gap

A frontier LLM queried cold about shopkunes.com returns: *"I do not have specific, verified information... It is not a widely recognized or documented e-commerce platform."* The model has zero awareness of Kunes Auto Group's 40+ locations, 23,000+ annual sales, 2,000+ employees, or its 52 "Best Dealerships to Work For" awards. The site itself is a substantial regional dealership group with press mentions in Bloomberg, CNBC, AARP, Automotive News, and Car Dealership Guy — yet none of that external signal has penetrated the model's training data. The gap between the brand's real footprint (a 40-store, multi-state operation founded in 1996) and its AI invisibility is near-total.

Schema Posture

Every page carries AutoDealer and WebSite JSON-LD with SearchAction — a strong foundation. However, the schema contains critical data-quality issues: the streetAddress field is literally "undefined" across all pages, the telephone field is empty, and openingHoursSpecification has malformed nested arrays (double-wrapped [[]] for departments). Individual vehicle inventory pages lack Product or Vehicle schema entirely — the sitemap lists 8,310 URLs, but none carry item-level structured data that would power rich AI answers about specific cars.

External Signals

The "In the News" page documents 20+ press mentions including Bloomberg (hybrid vehicle demand feature), CNBC (new vs. used car pricing), AARP, Automotive News, and PR Newswire. The brand sponsors Summerfest's Lasso Lounge and has a $100,000 United Way partnership. Yet web searches for "Kunes Auto Group" return zero indexed results in DuckDuckGo — suggesting the site's external SEO footprint is anomalously weak for a business of this scale. The domain has no Wayback Machine snapshots, no MX mail records, and no CDN/WAF layer beyond Google Frontend.

Findings

1. Zero cold LLM awareness despite being a major regional auto group High

   A frontier LLM queried cold about shopkunes.com returns no specific information, despite the group operating 40+ locations, selling 23,000+ vehicles annually, and having 2,000+ employees. The brand's real footprint is completely absent from AI training data.

   What to change: Increase external signals by securing press coverage in major outlets and ensuring the site is indexed by search engines. Consider submitting the site to AI training data sources.

2. llms.txt is a massive directory dump with generic descriptions Medium

   The llms.txt file is 270 KB and lists thousands of inventory URLs, but every entry uses the same generic description 'New quality vehicles in Delavan,' making it a low-signal resource for AI crawlers rather than a curated, high-value summary.

   What to change: Rewrite llms.txt to include unique, descriptive summaries for each page, highlighting key differentiators like location, vehicle types, and awards.

3. AutoDealer schema contains 'undefined' streetAddress High

   The streetAddress field in the AutoDealer JSON-LD is literally set to 'undefined' across all pages, which will cause AI crawlers to ignore or misinterpret the location data.

   What to change: Populate the streetAddress field with the correct physical address in the JSON-LD schema on every page.

4. AutoDealer schema has empty telephone field Medium

   The telephone field in the AutoDealer JSON-LD is empty, missing a critical piece of contact information that AI assistants rely on for local business queries.

   What to change: Add the correct phone number to the telephone field in the JSON-LD schema.

5. OpeningHoursSpecification has malformed nested arrays Medium

   The openingHoursSpecification in the schema contains double-wrapped arrays (e.g., [[]] for departments), which may cause parsing errors for AI crawlers.

   What to change: Fix the nested array structure in openingHoursSpecification to use single-level arrays for each department.

6. Inventory pages lack Product or Vehicle schema High

   Individual vehicle inventory pages do not carry Product or Vehicle structured data, preventing AI assistants from surfacing specific car details like make, model, price, and features in rich answers.

   What to change: Add Product or Vehicle schema to each inventory page with fields for make, model, year, price, mileage, and VIN.

7. Zero indexed results in DuckDuckGo for a major regional brand High

   Web searches for 'Kunes Auto Group' return zero indexed results in DuckDuckGo, indicating an anomalously weak external SEO footprint for a business of this scale.

   What to change: Improve SEO by building backlinks, optimizing on-page content, and ensuring the site is properly indexed by search engines.

8. No Wayback Machine snapshots for the domain Low

   The domain has no snapshots in the Wayback Machine, suggesting limited historical web presence or recent domain changes.

   What to change: Ensure the site is regularly crawled by the Internet Archive by maintaining a stable sitemap and robots.txt.

9. No MX mail records configured Low

   The domain lacks MX mail records, which may affect email deliverability and professional credibility.

   What to change: Configure MX records for the domain to enable email services.

What's working

• All major AI crawlers receive full, unblocked access — Every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, etc.) receives a 200 response with identical content to a browser, with no UA-based blocking.
• Next.js SSR serves server-rendered HTML to all bots — The site runs on Next.js SSR (RideMotive platform) via Google Frontend, ensuring that all crawlers receive fully rendered HTML content rather than empty JavaScript shells.
• AutoDealer and WebSite JSON-LD present on all pages — Every page carries AutoDealer and WebSite JSON-LD with SearchAction, providing a strong foundational schema structure for AI crawlers.
• llms.txt file is published and accessible — An llms.txt file exists at the standard location, signaling to AI crawlers that the site is aware of AI visibility best practices.
• Press mentions in Bloomberg, CNBC, and other major outlets — The 'In the News' page documents 20+ press mentions including Bloomberg, CNBC, AARP, and Automotive News, providing a rich source of external signals that could be leveraged for AI visibility.
• Philanthropy page contains substantial, unique content — The philanthropy page has 5,864 words of detailed community involvement content, which is valuable for AI crawlers seeking to understand the brand's values and local impact.