amfmtreatment.com — AI Site Grade

Analysis

The llms.txt is 676KB — larger than most entire websites — and the robots.txt has zero AI-bot directives, creating a paradoxical posture where AI crawlers get full access to a massive AI-friendly content map but no guidance on what to prioritize.

Crawler Access

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Applebot-Extended — receives a full 200 response with identical byte size (673,592 bytes) to a browser baseline. Only Bytespider gets a 403 block from Cloudflare. The robots.txt contains no AI-bot-specific rules whatsoever; the wildcard * rule only disallows /wp-admin/. The site runs on WP Engine behind Cloudflare DNS (NS: chan.ns.cloudflare.com, trevor.ns.cloudflare.com). DNS TXT records confirm explicit verification tokens for OpenAI (openai-domain-verification=dv-AbZdo1SJBA0BDP8fm7cei8xA), Anthropic (anthropic-domain-verification-06nrhk=...), and Apple — indicating deliberate AI-ecosystem enrollment.

The llms.txt Paradox

The site publishes an /llms.txt that is 676,507 bytes — an extraordinary size that includes hundreds of blog posts, condition pages, and location pages in a flat list. A dedicated llms-sitemap.xml exists solely to point to this single file. However, the robots.txt gives AI crawlers no prioritization signals, no crawl-delay for AI bots specifically, and no indication of which of the ~970+ sitemap URLs matter most. The llms.txt itself is a raw dump of every page with no hierarchy, no section grouping beyond a flat "Pages" list, and no "Optional" or "Preferred" sections as the llms.txt standard recommends.

Cold-Knowledge Gap

The LLM's cold knowledge describes AMFM as operating in California, Washington, and Virginia with a focus on co-occurring substance use issues, and mentions a founder "Dr. James E. D. M." with a military background. The actual site says nothing about a founder with that name, does not mention Washington as a primary state (though a Washington location appears in the locations list), and frames itself around adult residential mental health — not co-occurring substance use. The site's origin story centers on a family who lost their son Michael to mental illness, not a psychiatrist-founder. The LLM also recalls "mixed reviews" and "luxury/premium" positioning; the site itself claims "87% of clients would recommend AMFM" from a 2025 outcomes study but does not surface independent third-party reviews prominently.

Schema Posture

The homepage carries rich JSON-LD including MedicalWebPage, MedicalOrganization with hasCertification (listing CDSS/CCL, Joint Commission, NAATP), and FAQPage on the About page. However, the schema contains duplicate @graph blocks — one set with WebSite/WebPage/Article and another with MedicalWebPage — both pointing at the same homepage URL. The WebPage name in the second block reads "New Home 2025 – New Test", a staging/test label that leaked into production schema. Individual location pages lack LocalBusiness or MedicalBusiness schema with address/geo coordinates.

Broken Internal Navigation

Several key navigation links silently redirect to the homepage: /about/what-makes-us-different/ and /meet-the-team/ both return 200 but serve the homepage content with the homepage canonical URL. The site navigation lists "What Makes Us Different" and "Meet the Team" as distinct pages, but these URLs are non-functional — AI crawlers following those links land on duplicate homepage content with no unique information.

Findings

1. llms.txt lacks hierarchy and prioritization sections Medium

   The llms.txt is a flat 676KB dump of all pages with no 'Optional' or 'Preferred' sections as recommended by the standard, giving AI crawlers no guidance on which content matters most.

   What to change: Restructure llms.txt into 'Preferred' and 'Optional' sections, prioritizing core pages like locations, treatments, and outcomes study.

2. Robots.txt has no AI-bot-specific rules Medium

   The robots.txt only disallows /wp-admin/ for all user-agents, with no crawl-delay or allow/disallow rules for GPTBot, ClaudeBot, or other AI crawlers, despite the site's AI enrollment.

   What to change: Add specific directives for AI crawlers, such as a crawl-delay and allow rules for key sections, to signal prioritization.

3. Staging test label 'New Home 2025 – New Test' in production schema High

   The JSON-LD on the homepage contains a WebPage name reading 'New Home 2025 – New Test', indicating a staging/test label leaked into production, which can confuse AI crawlers and reduce trust.

   What to change: Remove the duplicate @graph block containing the staging label and ensure only clean, production-ready schema is served.

4. Duplicate @graph blocks in homepage JSON-LD Medium

   The homepage schema contains two @graph blocks with overlapping types (WebSite, WebPage, MedicalWebPage) pointing to the same URL, which may cause validation issues and dilute semantic signals.

   What to change: Merge the two @graph blocks into a single coherent graph, removing redundant entries.

5. Location pages lack LocalBusiness or MedicalBusiness schema High

   Individual location pages do not include LocalBusiness or MedicalBusiness schema with address, geo coordinates, or opening hours, reducing their visibility in AI-generated local results.

   What to change: Add LocalBusiness or MedicalBusiness schema to each location page with accurate address, phone, and geo coordinates.

6. Key navigation pages silently redirect to homepage High

   The URLs /about/what-makes-us-different/ and /meet-the-team/ return 200 but serve the homepage content with the homepage canonical URL, making them non-functional for AI crawlers and users.

   What to change: Either restore unique content to these pages or remove them from navigation and set proper 301 redirects to relevant pages.

7. LLM cold knowledge contradicts site content on founder and focus Medium

   LLMs recall a founder 'Dr. James E. D. M.' with military background and a co-occurring substance use focus, but the site's origin story centers on a family who lost their son Michael, and the focus is adult residential mental health.

   What to change: Ensure the site's About page and other key pages clearly state the founder story and treatment focus to align with LLM knowledge or correct misconceptions.

8. Independent third-party reviews not prominently surfaced Low

   The site claims high satisfaction from an internal outcomes study but does not prominently display independent reviews from platforms like Google or Yelp, which LLMs may rely on for reputation signals.

   What to change: Add a testimonials or reviews section that aggregates third-party reviews or links to external review profiles.

What's working

• All major AI crawlers receive full access with 200 responses — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others get full 200 responses with identical content to browsers, ensuring AI crawlers can index the site without restrictions.
• Domain verification tokens for OpenAI, Anthropic, and Apple — DNS TXT records include verification tokens for OpenAI, Anthropic, and Apple, indicating deliberate enrollment in AI ecosystems for enhanced crawling and indexing.
• Massive llms.txt with 676KB of content — The site publishes an llms.txt file containing hundreds of pages, providing AI crawlers with a comprehensive content map for efficient discovery.
• Rich JSON-LD schema on homepage with MedicalWebPage and certifications — The homepage includes MedicalWebPage, MedicalOrganization with hasCertification (CDSS/CCL, Joint Commission, NAATP), and FAQPage, providing strong semantic signals.
• 2025 Outcomes Study Report with client satisfaction data — The site publishes a detailed outcomes study report claiming 87% of clients would recommend AMFM, providing authoritative data for AI citation.
• Sitemap with 80+ URLs covering key sections — The sitemap includes 80 URLs covering locations, treatments, blog, and other key pages, aiding crawler discovery.
• Dedicated Licenses & Accreditations page with detailed credentials — The site has a page listing licenses and accreditations (CDSS, Joint Commission, NAATP), building trust with AI crawlers and users.