carefertility.com — AI Site Grade

Analysis

The cold-knowledge model claims a 2023 data breach — but no evidence of this exists on the site, in search results, or in press coverage, making it a hallucinated reputational liability that AI engines may surface alongside the brand.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, Bytespider, anthropic-ai — receive a 200 with full HTML content identical to browser baseline (~204KB). No UA-based blocking, no Cloudflare challenge, no JS shell. The site runs on HubSpot behind Cloudflare (strict-transport-security enabled, no X-Frame-Options). The robots.txt is a generic HubSpot template with no AI-bot directives whatsoever — it only blocks HubSpot CMS preview/management paths. llms.txt returns a 404 (HubSpot 404 page, 86KB of CSS/JS). The sitemap.xml exists with 674 URLs and is well-structured.

Cold-Knowledge Gap

The LLM prior knows Care Fertility as a UK fertility network founded in 1997 by Dr. Simon Fishel (part of the team behind Louise Brown, the first IVF baby) and co-founder Dr. John Webster. It mentions "Care Maps" treatment plans and "EmbryoGlue." The site's "Our Story" page confirms the founding narrative (Simon Fishel, John Webster, Bourn Hall lineage) but never mentions EmbryoGlue — the product the model associates most distinctively with the brand. The site instead emphasizes Caremaps-AI (AI embryo selection) and the Salve treatment companion app, which the model knows nothing about. The model also claims a 2023 data breach involving patient records — no evidence of this exists anywhere on the site, in web search results, or in press coverage, making this a hallucinated reputational risk that could surface in AI-generated summaries.

Schema Posture

The site uses Organization schema on every page (name, address, phone, language) and BreadcrumbList on deeper pages. The clinics page uses MedicalClinic schema with geo-coordinates, telephone, and parentOrganization — strong local SEO. The blog uses BlogPosting schema with author and headline. However, no FAQPage schema exists anywhere despite the site having a /help/ section with hundreds of question-answer pages (e.g., "Can I get NHS funding?", "What are the steps of IVF?"). These pages are plain HTML with no structured markup. The homepage has no FAQ schema despite a "questions we hear the most" section. No Product or MedicalBusiness schema for treatment packages. No VideoObject schema on the IVF page despite embedded video content (though a VideoObject was found in the raw JSON-LD for the IVF page with clips).

External Signals

The site prominently links to Trustpilot ("Rated excellent") and HFEA, Fertility Network UK, and NHS England as authority signals. DNS records show anthropic-domain-verification and apple-domain-verification TXT records — the brand has proactively verified ownership for AI crawlers. The blog is active (posts dated May 2026, April 2026) and covers timely topics (AI in fertility, Black maternal health, egg freezing trends). The news section includes a press release debunking "Fertox" social media detox myths, positioning the brand as an evidence-based authority. No external press coverage of a data breach was found.

Content Architecture Surprise

The /help/ section contains hundreds of short-form Q&A pages (each ~200-250 words) organized by topic categories (Paying for Treatment, Fertility Treatments, Support & Wellbeing). These are effectively an FAQ knowledge base but lack FAQPage schema, comparison tables, or any structured answer format. The /success/ page is notably thin (260 words, no actual success rate numbers or HFEA data displayed — just links to "Read More"). The homepage claims "some of the highest success rates of any UK clinic" but the success page itself is a landing page with no data tables. The /clinics/ page has excellent MedicalClinic schema for 22+ locations but renders only 52 visible words of text — the clinic list is loaded via JavaScript/map interface, which may be invisible to some crawlers despite the JSON-LD being present.

Findings

1. Hundreds of Q&A pages lack FAQPage schema High

   The /help/ section contains hundreds of short-form Q&A pages (e.g., 'Can I get NHS funding?', 'What are the steps of IVF?') that are plain HTML with no FAQPage structured data. This prevents AI crawlers from extracting answers directly for featured snippets or voice search.

   What to change: Add FAQPage schema to all /help/ pages, marking each question-answer pair as a mainEntity item.

2. LLM knowledge hallucinates a 2023 data breach High

   The cold-knowledge model claims a 2023 data breach involving patient records, but no evidence exists on the site, in search results, or in press coverage. This hallucinated fact could surface in AI-generated summaries, damaging brand trust.

   What to change: Publish a dedicated page or press release addressing the false breach claim, and use schema markup (e.g., FAQPage or Article) to provide authoritative context that AI crawlers can index.

3. llms.txt returns 404 Medium

   The llms.txt file is missing, returning a HubSpot 404 page. This file helps AI crawlers discover key content and context; its absence means the site misses an opportunity to guide AI summarization.

   What to change: Create an llms.txt file listing important pages (e.g., treatments, clinics, help section) and a brief brand description.

4. Success rates page lacks data tables Medium

   The /success/ page is only 260 words with no actual success rate numbers or HFEA data displayed; it merely links to 'Read More'. This undermines the homepage claim of 'some of the highest success rates' and provides no structured data for AI crawlers.

   What to change: Add a data table with HFEA success rates and mark it up with Dataset or MedicalBusiness schema.

5. Clinic list rendered via JavaScript Medium

   The /clinics/ page has only 52 visible words; the clinic list is loaded via a JavaScript map interface. While JSON-LD is present, some crawlers may not execute JS, making the list invisible to them.

   What to change: Include a static HTML list of clinics as a fallback for non-JS crawlers.

6. No Product or MedicalBusiness schema for treatments Medium

   Treatment pages (e.g., IVF) lack Product or MedicalBusiness schema, which could help AI crawlers understand pricing, packages, and services. The IVF page has 4141 words but no structured data for treatment offerings.

   What to change: Add MedicalBusiness or Product schema to treatment pages with offers, price, and description.

7. Embedded video lacks VideoObject schema Low

   The IVF page contains embedded video content but no VideoObject schema was found in the raw JSON-LD (though a VideoObject was noted for clips). This reduces the chance of video appearing in AI search results.

   What to change: Add VideoObject schema to all pages with embedded video, including description, thumbnail, and content URL.

8. Homepage FAQ section lacks schema Low

   The homepage has a 'questions we hear the most' section but no FAQPage schema, missing an opportunity for AI crawlers to extract Q&A content directly.

   What to change: Add FAQPage schema to the homepage FAQ section.

What's working

• All major AI crawlers receive full HTML content — All tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) receive a 200 with full HTML content identical to browser baseline. No UA-based blocking or Cloudflare challenges are present.
• Organization schema on every page — Every page includes Organization schema with name, address, phone, and language, providing consistent brand identity to AI crawlers.
• MedicalClinic schema on clinics page with geo-coordinates — The /clinics/ page uses MedicalClinic schema with geo-coordinates, telephone, and parentOrganization for 22+ locations, boosting local SEO and AI visibility.
• BlogPosting schema on blog articles — Blog articles use BlogPosting schema with author and headline, helping AI crawlers understand content structure.
• Domain verification TXT records for AI crawlers — DNS records include anthropic-domain-verification and apple-domain-verification TXT records, proactively verifying ownership for AI crawlers.
• Active blog and news section with timely content — The blog is regularly updated (posts dated May 2026, April 2026) covering AI in fertility, Black maternal health, and egg freezing trends. The news section includes evidence-based content debunking social media myths.
• Links to Trustpilot, HFEA, and NHS England — The site prominently links to Trustpilot, HFEA, Fertility Network UK, and NHS England, providing external authority signals that AI crawlers can use to assess credibility.
• Well-structured sitemap with 674 URLs — The sitemap.xml exists with 674 URLs and is well-structured, helping crawlers discover all pages efficiently.