buylewis.com — AI Site Grade

Analysis

I have enough information to write the audit. Let me compile the key findings.

The site is buylewis.com — Lewis Automotive Group, a multi-franchise auto dealer in western Kansas (Hays, Dodge City, Garden City, Liberal). The critical finding: the site is completely blocked to all browsers and most AI crawlers at the Akamai edge (403), but the anthropic-ai bot user-agent gets through to an nginx backend with full content. Meanwhile, ClaudeBot is blocked. The llms.txt exists and is massive (399KB). No external search results exist for this brand. The cold LLM knowledge is empty.

Let me write the audit.

AI-Visibility Audit: buylewis.com (Lewis Automotive Group)

The site is fully walled off from every browser and nearly every AI crawler by an Akamai edge block, yet the anthropic-ai user-agent alone passes through to a live nginx backend serving 340KB of real content — a backdoor that creates a single-crawler monoculture for AI discoverability.

Crawler Access

Every tested user-agent — GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Bytespider, Applebot-Extended, and even a standard browser — receives a 403 from AkamaiGHost on the homepage, robots.txt, and llms.txt. The sole exception is anthropic-ai (the generic Anthropic research crawler), which gets a 200 from nginx with the full HTML page. The robots.txt (only visible to anthropic-ai) explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, Claude-User, and Claude-SearchBot from /api/, /apis/, and /pixall/ paths — but the Akamai block makes those rules moot since those bots never reach the file. The llms.txt exists and is 399KB — an unusually large, auto-generated listing of every page on the site — but only the anthropic-ai crawler can read it.

Cold-Knowledge Gap

A frontier LLM queried cold about buylewis.com returns zero prior knowledge: no awareness of the brand, its locations, its franchises, or its market position. This is a total knowledge vacuum despite the site being a multi-franchise dealer group (Toyota, Ford, Chevrolet, Chrysler, Dodge, Jeep, Ram, Nissan, Buick, GMC) with physical locations in four western Kansas cities. The gap between the site's actual content and what AI models know about it is absolute.

Content & Schema Posture

The homepage and inventory pages are built on the Dealer.com (DDC) platform — a common automotive CMS. The site carries meta name="robots" content="index, follow" on key pages, but those directives are irrelevant since only one crawler can reach them. No JSON-LD structured data was detected in the page head. The site uses heavy JavaScript (prefetch bundles for inventory facets, filters, listings) — a JS-rendering risk for any crawler that does execute scripts. Key pages like /about-us.htm and /contact-us.htm return 404 errors (the actual about page lives at /dealership/about.htm), meaning standard navigation paths are broken.

External Signals

Web search returns zero indexed results for buylewis.com, "Lewis Auto Group," or "Lewis Automotive Group" — no organic listings, no reviews, no Reddit threads, no press mentions. The domain has no discoverable external footprint. DNS records show Cloudflare nameservers, Akamai edge IPs, and Microsoft 365 mail — a split infrastructure where Akamai enforces the blanket block while nginx serves content only to the one permitted crawler.

Findings

1. Akamai edge blocks all browsers and most AI crawlers with 403 High

   Every tested user-agent except anthropic-ai receives a 403 from AkamaiGHost on the homepage, robots.txt, and llms.txt. This includes GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Bytespider, Applebot-Extended, and standard browsers.

   What to change: Remove the blanket Akamai block for AI crawler user-agents, or configure Akamai to allow specific AI crawlers (e.g., GPTBot, ClaudeBot) to reach the nginx backend.

2. ClaudeBot blocked but anthropic-ai allowed, creating a single-crawler monoculture High

   The anthropic-ai crawler (generic research bot) gets a 200 from nginx with full HTML, while ClaudeBot (Anthropic's production crawler) receives a 403. This inconsistency limits AI discoverability to only one crawler.

   What to change: Allow ClaudeBot and other major AI crawlers (GPTBot, OAI-SearchBot, etc.) through the Akamai edge to the nginx backend.

3. Robots.txt only visible to anthropic-ai, disallows other AI bots from certain paths Medium

   The robots.txt file is only accessible to the anthropic-ai crawler; all other bots get a 403. It explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, Claude-User, and Claude-SearchBot from /api/, /apis/, and /pixall/ paths, but those rules are moot since those bots never reach the file.

   What to change: Make robots.txt publicly accessible and remove disallow rules for AI crawlers on non-sensitive paths.

4. Llms.txt only accessible to anthropic-ai, 399KB auto-generated listing Medium

   The llms.txt file is 399KB and contains an auto-generated listing of every page on the site, but only the anthropic-ai crawler can read it. Other AI crawlers receive a 403.

   What to change: Make llms.txt publicly accessible and consider trimming it to a curated, concise summary of the site's key pages.

5. Zero external search results for brand or domain High

   Web searches for buylewis.com, 'Lewis Auto Group', and 'Lewis Automotive Group' return zero indexed results. No organic listings, reviews, or press mentions exist.

   What to change: Implement SEO best practices to get pages indexed: ensure crawler access, submit sitemap to search engines, and build external backlinks.

6. Cold LLM knowledge vacuum for Lewis Automotive Group High

   A frontier LLM queried cold about buylewis.com returns zero prior knowledge: no awareness of the brand, its locations, franchises, or market position.

   What to change: Improve AI visibility by allowing major AI crawlers access and ensuring structured data is present on key pages.

7. About and contact pages return 404 errors Medium

   Standard navigation paths /about-us.htm and /contact-us.htm return 404 errors. The actual about page lives at /dealership/about.htm, meaning standard navigation paths are broken.

   What to change: Fix the broken URLs by redirecting /about-us.htm and /contact-us.htm to the correct pages or restoring the pages at those paths.

8. No JSON-LD structured data detected on key pages Medium

   No JSON-LD structured data was detected in the page head of the homepage or inventory pages. This limits AI models' ability to extract entity information.

   What to change: Add JSON-LD structured data for LocalBusiness, AutoDealer, and VehicleOffer schemas on relevant pages.

9. Heavy JavaScript rendering risk for AI crawlers Medium

   The site uses heavy JavaScript (prefetch bundles for inventory facets, filters, listings) — a JS-rendering risk for any crawler that does not execute scripts.

   What to change: Ensure critical content is server-side rendered or included in the initial HTML payload to be accessible to crawlers that do not execute JavaScript.

10. Sitemap inaccessible to most bots due to Akamai block High

    The sitemap.xml returns 403 for most bots; only anthropic-ai can access it. This prevents search engines and AI crawlers from discovering the site's URL structure.

    What to change: Make sitemap.xml publicly accessible and ensure it is listed in robots.txt.

What's working

• Anthropic-ai crawler receives full HTML content with 200 status — The anthropic-ai user-agent successfully retrieves the full homepage (340KB) and other pages with a 200 status, indicating the nginx backend serves complete content.
• Llms.txt exists and is accessible to anthropic-ai — The site has an llms.txt file (399KB) that lists all pages, providing a comprehensive resource for AI crawlers that can access it.
• Robots.txt disallows AI bots from sensitive paths — The robots.txt explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, Claude-User, and Claude-SearchBot from /api/, /apis/, and /pixall/ paths, which is a good practice to prevent crawling of non-public endpoints.
• Meta robots index, follow on key pages — Key pages carry meta name='robots' content='index, follow', indicating the site intends for these pages to be indexed.
• Nginx backend serves full content to permitted crawler — The nginx backend returns 200 with full HTML for the anthropic-ai crawler, showing the site has a functional content delivery path.