accurx.com — AI Site Grade

Analysis

Accurx has no structured data of any kind across its entire site — zero JSON-LD, zero schema.org markup — despite being a health-tech platform used by 98% of GP practices and 68% of NHS trusts, making it invisible to knowledge-graph-driven AI features like Google's AI Overviews, ChatGPT's structured citations, and Perplexity's rich answer panels.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Applebot-Extended — receive a full 200 response with identical content to a browser baseline (72,719 bytes). Bytespider (ByteDance) is the sole bot blocked, returning a 403 from Cloudflare. The robots.txt is minimal: a single User-agent: * rule disallowing only /faq/ and /5-point-prescription, with no AI-bot-specific directives. No llms.txt exists (404). The site is built on Webflow and served behind Cloudflare with Azure DNS. Content is server-rendered HTML, not a JS shell — AI crawlers can read every page fully.

Cold-Knowledge Gap

The LLM prior knows Accurx as a UK health-tech company founded in 2016 by Dr. Jacob Haddad and Dr. Laurence Bargery, focused on NHS communication tools, with over £100M in funding and a "unicorn" reputation. It also recalls 2023 criticism from GPs and patient groups over data privacy and triage workload concerns — a reputational signal the site itself never addresses. The site's homepage and product pages make no mention of the founders by name, no mention of funding rounds or valuation, and contain no trust signals (security certifications, ISO badges, data-center compliance details) that would counter the privacy criticism. The site presents Accurx as an "easy-to-use platform" but the cold model knows it as a well-capitalized, occasionally controversial health-tech unicorn — two different brands.

Schema Posture

Zero structured data exists on any page examined. The homepage, /scribe, /patient-triage, /primary-care, /who-we-are, /blog, /press, and /case-studies all return no JSON-LD, no @graph, no schema.org types of any kind. There is no Organization schema, no Product schema for Scribe or Patient Triage, no FAQPage schema on the FAQ-heavy triage page, no Article schema on blog posts, no BreadcrumbList, no WebSite schema with potentialAction for search. For a platform that claims "98% of GP practices choose us" and "Used in 68% of NHS trusts," the absence of Organization schema with those statistics as structured claims is a significant gap for AI answer engines that pull from knowledge graphs.

External Signals

The press page lists coverage from Forbes, Financial Times, Digital Health, UKAuthority, and multiple NHS-focused outlets. Recent press (March–May 2026) heavily features the Accurx Scribe / Ambient Voice Technology rollout across University Hospitals of Leicester and Northamptonshire — a £1.9M contract covering 10,000 clinicians. This is the site's strongest current narrative, yet the homepage buries it under generic messaging. The blog contains substantive, well-written posts (e.g., the AVT rollout post with pilot data: 8 minutes saved per patient, same-day letters vs. 2-3 days), but none of this data is surfaced in structured format. No Reddit threads or independent review sites appeared in search results — the site's external reputation is almost entirely mediated through NHS trade press.

Findings

1. Zero structured data across the entire site High

   No JSON-LD or schema.org markup exists on any page examined, including homepage, product pages, blog, press, and case studies. This prevents AI systems from extracting entity relationships, statistics, or product details for knowledge graphs and rich answer panels.

   What to change: Add Organization schema to the homepage with name, description, founding date, founders, funding, and key statistics. Add Product schema for Scribe and Patient Triage. Add Article schema to blog posts and FAQPage schema to the triage page.

2. Cold-knowledge gap between brand and public perception High

   The LLM prior knows Accurx as a well-funded unicorn with past privacy criticism, but the site omits founders, funding, and trust signals like security certifications. This mismatch undermines AI-generated summaries that may surface negative context without counterbalancing positive signals.

   What to change: Add an 'About' section with founder names, funding milestones, and trust badges (ISO, Cyber Essentials) to the homepage and /who-we-are page. Include these in Organization schema.

3. No llms.txt file for AI crawler guidance Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a curated summary of key pages and content for retrieval-augmented generation.

   What to change: Create an llms.txt file listing the most important pages (homepage, product pages, press, case studies) with brief descriptions.

4. robots.txt blocks FAQ and 5-point-prescription pages Medium

   The robots.txt disallows /faq/ and /5-point-prescription for all user-agents, preventing AI crawlers from indexing potentially valuable content that could answer common questions.

   What to change: Remove the disallow rules for /faq/ and /5-point-prescription, or allow AI bots specifically.

5. Bytespider (ByteDance) blocked by Cloudflare Low

   Bytespider receives a 403 response, blocking content from being indexed by ByteDance's AI systems (e.g., Doubao). While not a top-tier AI crawler, this reduces potential visibility in Asian markets.

   What to change: Allow Bytespider access by adjusting Cloudflare WAF rules.

6. Homepage buries strongest press narrative Medium

   The homepage uses generic messaging ('The simple way to communicate about patient care') instead of highlighting the recent £1.9M AVT rollout covered by Forbes and Financial Times, which is the site's most compelling current story.

   What to change: Feature the AVT rollout and press mentions prominently on the homepage, with structured data to highlight the statistics.

7. No FAQ schema on patient triage page despite FAQ content Medium

   The /patient-triage page contains extensive FAQ-style content but lacks FAQPage schema, missing an opportunity to appear in AI-generated FAQ snippets.

   What to change: Add FAQPage schema with questions and answers from the triage page content.

8. No Article schema on blog posts Medium

   Blog posts, including the AVT rollout article with concrete data, lack Article schema, reducing their chance of being featured in Google's Top Stories or AI-driven news summaries.

   What to change: Add Article schema to all blog posts with headline, datePublished, author, and image.

9. No BreadcrumbList schema on any page Low

   The site does not implement BreadcrumbList schema, which helps AI systems understand site hierarchy and improves navigation in search results.

   What to change: Add BreadcrumbList schema to all pages reflecting the site's navigation structure.

10. No Product schema for Scribe Medium

    The Scribe product page lacks Product schema, missing the chance to provide AI systems with structured details about the AI scribe tool, its features, and pricing.

    What to change: Add Product schema with name, description, brand, and offers (if applicable).

11. No WebSite schema with search action Low

    The site does not implement WebSite schema with potentialAction for search, which would enable AI systems to perform site searches directly.

    What to change: Add WebSite schema with SearchAction pointing to a search endpoint.

12. No external review presence on Reddit or independent sites Low

    Web searches for Accurx on Reddit and general review sites returned zero results, indicating a lack of organic third-party discussion that AI systems often surface.

    What to change: Encourage user reviews on platforms like G2 or Capterra, and engage in relevant Reddit communities.

What's working

• All major AI crawlers allowed with full content access — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others receive 200 responses with identical content to browser views, ensuring AI systems can read all pages.
• Server-rendered HTML content accessible to crawlers — The site is built on Webflow with server-rendered HTML, not a JavaScript shell, so AI crawlers can parse all content without needing to execute JavaScript.
• Strong press coverage from major outlets — The press page lists coverage from Forbes, Financial Times, Digital Health, and others, providing authoritative external signals that AI systems can cite.
• Detailed case studies with concrete data — Case studies and blog posts contain specific metrics (e.g., 8 minutes saved per patient, same-day letters) that are valuable for AI-generated answers, though not in structured format.
• Comprehensive sitemap with 80 URLs — The sitemap contains 80 URLs, ensuring good discoverability of pages by search engines and AI crawlers.
• No JavaScript shell pages — All pages tested return full HTML content without requiring JavaScript rendering, making them fully accessible to AI crawlers.
• Clean robots.txt with no AI-bot blocking — The robots.txt does not block any AI crawlers, allowing full access to all pages except /faq/ and /5-point-prescription.
• Fast and reliable hosting with Cloudflare and Azure DNS — The site uses Cloudflare for CDN and security, and Azure DNS for reliability, ensuring fast load times and high availability for crawlers.