hotchkissinsurance.com — AI Site Grade

Analysis

Hotchkiss Insurance runs an llms.txt file — a rare, advanced AI-readiness signal — yet simultaneously blocks ClaudeBot and Bytespider at the server level with 403s, creating a contradictory posture where the site invites AI consumption via one mechanism while actively rejecting it via another.

Crawler Access

The site is hosted on a single DigitalOcean nginx server (167.99.229.103) with no CDN or WAF. compare_bot_access reveals a split pattern: GPTBot, Google-Extended, OAI-SearchBot, PerplexityBot, ChatGPT-User, and anthropic-ai all receive a full 200 with identical 46KB HTML payloads. ClaudeBot and Bytespider get 403 responses. The robots.txt contains no AI-bot directives at all — only generic disallows for /cpresources/, /vendor/, /.env, and /cache/. The absence of any CDN means every AI crawler hits the origin server directly, and the lack of X-Robots-Tag or security headers leaves the site exposed to straightforward scraping.

llms.txt — A Surprising Asset

The site has a well-structured llms.txt at /llms.txt (8,252 bytes) — a convention most sites in insurance do not implement. It includes a brand summary, core service links, industry solutions, and company information. This is the strongest AI-readiness signal on the domain. However, the llms.txt lists URLs that ClaudeBot cannot access (403), creating a discoverability dead end for Anthropic's crawler.

Cold-Knowledge Gap

The LLM's cold knowledge about Hotchkiss Insurance contains a significant factual error: it states the agency was founded in 1969. The site consistently states 1975 (founded by Doug Hotchkiss). The model also mentions a "Best Places to Work" recognition from the Houston Business Journal in 2023 — a signal that does not appear anywhere on the site's own pages. The site prominently features a 2024 Top 100 Brokers of U.S. Business ranking (debut at #93), which the model does not know about. The model also references "transportation" as a served industry, but the site lists 13 industries and transportation is not among them.

Schema and Content Posture

Every page carries consistent LocalBusiness JSON-LD with full NAP (name, address, phone), opening hours, and social profiles. The homepage schema incorrectly uses HealthInsurancePlan as its primary @type — a mismatch for a multi-line insurance brokerage. No FAQPage, Product, Article, or HowTo schema appears anywhere, even on the blog/insights pages. The insights section (at /insights-and-resources) contains dated articles with future timestamps (e.g., "5.18.2026"), suggesting a publishing system misconfiguration that could confuse crawlers about content freshness.

External Signals

Web searches returned zero indexed results for review sites, press mentions, or third-party coverage of Hotchkiss Insurance. The domain has no discoverable external reputation footprint — no Google Business Profile snippets, no Reddit threads, no news articles surfaced. The site's own social links point to Twitter (@hiallc), Facebook, Instagram, and LinkedIn, but none of these profiles appear in search results for the brand name. This near-total absence of external signals means AI models have almost no third-party corroboration to draw on, making the site's own content the sole source of truth — and any errors in that content (like the 1969 vs 1975 founding date) propagate unchallenged.

Findings

1. ClaudeBot and Bytespider receive 403 responses High

   ClaudeBot and Bytespider are blocked at the server level with 403 responses, while other AI crawlers like GPTBot and Google-Extended get full 200s. This contradicts the site's llms.txt invitation.

   What to change: Allow ClaudeBot and Bytespider access by removing the 403 rules in the nginx configuration.

2. Robots.txt lacks AI crawler directives Medium

   The robots.txt file contains no rules for AI bots, leaving crawler access entirely to server-level controls. This creates an opaque access policy.

   What to change: Add explicit allow/disallow directives for major AI crawlers in robots.txt.

3. llms.txt lists URLs that ClaudeBot cannot access High

   The llms.txt file includes URLs that return 403 for ClaudeBot, creating a discoverability dead end for Anthropic's crawler.

   What to change: Ensure all URLs referenced in llms.txt are accessible to ClaudeBot.

4. LLM cold knowledge states incorrect founding year (1969 vs 1975) High

   The LLM's prior knowledge claims Hotchkiss Insurance was founded in 1969, but the site consistently states 1975. This factual error can propagate in AI outputs.

   What to change: Add structured data (e.g., foundingDate in LocalBusiness schema) and reinforce the correct year across the site.

5. Homepage schema uses HealthInsurancePlan as primary type High

   The homepage JSON-LD uses HealthInsurancePlan as its primary @type, which is incorrect for a multi-line insurance brokerage. This can confuse AI crawlers about the business nature.

   What to change: Change the primary @type to InsuranceAgency or LocalBusiness.

6. No FAQPage, Article, or HowTo schema on any page Medium

   Despite having an insights section and service pages, the site uses no FAQPage, Article, or HowTo structured data, missing opportunities for rich AI snippets.

   What to change: Add FAQPage schema to FAQ content, Article schema to blog posts, and HowTo schema to service explainers.

7. Insights articles have future timestamps (e.g., 5.18.2026) Medium

   The insights section contains articles with dates in the future, likely due to a publishing system misconfiguration. This can confuse crawlers about content freshness.

   What to change: Correct the date system to display actual publication dates.

8. No external signals found in web searches High

   Web searches for reviews, press mentions, and third-party coverage returned zero results. The site has no discoverable external reputation footprint, leaving AI models with only the site's own content.

   What to change: Encourage customer reviews on Google and third-party sites, and pursue press mentions to build external signals.

9. No CDN or WAF in place Medium

   The site is hosted on a single DigitalOcean nginx server with no CDN or WAF, meaning all AI crawlers hit the origin directly and the site lacks security headers.

   What to change: Add a CDN (e.g., Cloudflare) and a WAF to improve performance and security.

10. LLM cold knowledge lacks 2024 Top 100 Brokers ranking Medium

    The site prominently features a 2024 Top 100 Brokers ranking, but the LLM's prior knowledge does not include this, missing a key credibility signal.

    What to change: Add the ranking to structured data and ensure it is mentioned in llms.txt.

What's working

• Well-structured llms.txt file — The site has an 8,252-byte llms.txt with brand summary, service links, and company info, a rare AI-readiness signal in the insurance industry.
• Consistent LocalBusiness JSON-LD with full NAP — Every page carries LocalBusiness schema with correct name, address, phone, opening hours, and social profiles, aiding local AI visibility.
• Most major AI crawlers receive 200 responses — GPTBot, Google-Extended, OAI-SearchBot, PerplexityBot, ChatGPT-User, and anthropic-ai all get full HTML content, enabling AI indexing.
• Sitemap with 8 URLs and index — A sitemap is available and includes 8 URLs with an index, helping crawlers discover key pages.
• Detailed service and industry pages — Pages like /services, /industries, and /employee-benefits provide substantial content (362-514 words) covering offerings and industries.
• Social media profiles linked on site — The site links to Twitter, Facebook, Instagram, and LinkedIn profiles, providing additional signals for AI models.