solace.health — AI Site Grade

Analysis

Solace.health: A $1B Startup That AI Models Have Never Heard Of

The cold LLM knowledge about Solace Health describes a Canadian telehealth platform — a completely different company — while the actual site is a U.S. patient-advocacy marketplace that just raised $130M at a $1 billion valuation in February 2026. This is not a minor gap; it is a total identity erasure in the AI knowledge base.

Crawler Access

The robots.txt explicitly disallows every major AI crawler — GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Bytespider — via Cloudflare-managed rules. However, compare_bot_access shows that all bots actually receive a 200 with full HTML content (175KB, identical to browser baseline). The disallow directives are present but not enforced at the server level. Bytespider is the sole exception, returning a 403. The /llms.txt URL does not exist — it redirects to the homepage (200, same content). The /sitemap.xml also redirects to the homepage rather than serving an actual sitemap, meaning no structured URL map exists for any crawler.

Schema and Content Posture

Zero JSON-LD schema of any type exists on any page examined — homepage, about, FAQ, outcomes, articles, specialties, states, payers, and the press release. No Organization, LocalBusiness, FAQPage, Article, MedicalBusiness, or WebSite schema is present. The FAQ page uses semantic heading structure (H2/H3) but no FAQPage markup. The homepage has two H1 tags, which is a structural issue. The site is built on Webflow (visible in DNS TXT records and response headers), which typically renders content server-side — no JS-shell risk was detected.

Cold-Knowledge Gap

The frontier LLM describes Solace Health as "a Canadian digital health platform that connects patients with healthcare providers for virtual care... covered by provincial health insurance." The actual Solace is a U.S. company founded in 2022 by Jeremy Gurewitz, connecting patients with registered nurse advocates covered by Medicare, with 2,000+ advocates, 200,000+ patients served, and a $1B valuation from IVP, Menlo Ventures, and others. The model knows nothing about the patient-advocacy model, the Medicare billing angle, the Series C, or the company's actual domain.

External Signals

External search results returned zero indexed mentions of Solace Health across news, Reddit, or review sites — despite the company claiming a Bloomberg feature, a $1B valuation, and 8,000+ reviews on the outcomes page. The newsroom page contains placeholder links (example.com/sports-highlights, example.com/sustainability-initiatives) alongside the real Bloomberg press release. The DNS records show an anthropic-domain-verification TXT token, indicating the company has proactively registered with Anthropic — yet ClaudeBot is disallowed in robots.txt.

Findings

1. AI models describe a different company entirely High

   Frontier LLMs describe Solace Health as a Canadian telehealth platform, not the U.S. patient-advocacy marketplace that raised $130M at a $1B valuation. This complete identity mismatch means AI-generated answers will be wrong.

   What to change: Publish structured data (Organization, WebSite) and a verified llms.txt file to correct the AI knowledge base. Submit the site to AI crawler directories and ensure crawlers are allowed.

2. Robots.txt disallows all major AI crawlers High

   The robots.txt explicitly disallows GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, and Bytespider. While most bots still receive a 200 response, the disallow directives signal to compliant crawlers to stay away, and Bytespider is blocked with a 403.

   What to change: Remove disallow directives for AI crawlers from robots.txt, or replace them with allow rules for key pages.

3. Sitemap.xml redirects to homepage instead of serving URLs High

   The sitemap.xml URL redirects to the homepage, providing no structured URL list to any crawler. This prevents discovery of all pages.

   What to change: Generate and serve a valid sitemap.xml listing all important pages.

4. LLMs.txt redirects to homepage instead of providing AI guidance High

   The /llms.txt URL redirects to the homepage, offering no structured information for AI crawlers to understand the site's content and purpose.

   What to change: Create an llms.txt file that describes the site, key pages, and how AI should reference it.

5. No JSON-LD schema on any page High

   Every page examined — homepage, about, FAQ, outcomes, articles, specialties, states, payers, press release — lacks any JSON-LD structured data. No Organization, LocalBusiness, FAQPage, Article, or MedicalBusiness schema is present.

   What to change: Add JSON-LD schema for Organization, WebSite, FAQPage, Article, and LocalBusiness on relevant pages.

6. No external citations found despite claimed press coverage High

   Web searches for Solace Health across news, Reddit, and review sites returned zero indexed results, despite the site claiming a Bloomberg feature, a $1B valuation, and 8,000+ reviews. This severely limits AI citation sources.

   What to change: Ensure press releases are published on high-authority news sites and that reviews are indexed. Consider submitting to relevant directories and encouraging backlinks.

7. Newsroom page contains placeholder links alongside real press release Medium

   The newsroom page lists placeholder URLs (example.com/sports-highlights, example.com/sustainability-initiatives) alongside the real Bloomberg press release, which may confuse crawlers and reduce credibility.

   What to change: Remove placeholder links from the newsroom page.

8. Homepage has two H1 tags Low

   The homepage contains two H1 elements, which is a structural HTML issue that can confuse crawlers about the page's primary topic.

   What to change: Ensure only one H1 tag per page.

9. Anthropic domain verification token present but ClaudeBot disallowed Medium

   DNS records include an anthropic-domain-verification TXT token, indicating proactive registration with Anthropic, yet robots.txt disallows ClaudeBot. This contradiction undermines the verification's purpose.

   What to change: Remove the ClaudeBot disallow from robots.txt to align with the verification token.

What's working

• Content is server-side rendered and accessible to crawlers — The site is built on Webflow, which renders content server-side. All pages return full HTML (175KB) to crawlers, avoiding the JS-shell problem common with SPAs.
• FAQ page uses semantic heading structure — The FAQ page organizes questions with H2 and H3 headings, which helps crawlers understand the content hierarchy even without FAQPage schema.
• Anthropic domain verification token is in place — The DNS includes an anthropic-domain-verification TXT record, showing the company has proactively registered with Anthropic to assert ownership and potentially influence AI training.
• Site has a broad set of informational pages — The site includes pages for specialties, states, payers, outcomes, articles, and a press release, providing a rich content base that could be leveraged for AI visibility.