sevetys.fr — AI Site Grade

Analysis

Sevetys.fr — AI-Visibility Audit

The cold LLM knowledge gap is absolute: a frontier model queried on "Sevetys" returns zero recall — no sector, no products, no reputation — despite the site representing a veterinary group with ~250 clinics, 2,000+ professionals, and over 1 million animals treated, founded in 2016 and headquartered in Paris.

Crawler Access

All major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended) receive HTTP 200 with full HTML content identical to browser baseline (~997 KB). The robots.txt uses a single User-Agent: * catch-all with six Disallow rules blocking authentication pages, URL-parameter search filters, and the /prendre-rendez-vous/* booking path — but no AI-bot-specific directives exist. No llms.txt exists (404). The site runs Next.js on Azure (IP 20.19.233.176) with strong security headers (HSTS, CSP, X-Frame-Options: DENY) but no CDN/WAF layer that would interfere with bots.

Content & Schema Posture

The homepage carries a single MedicalBusiness JSON-LD block with name, address, founding date, VAT ID, and social links — solid for entity recognition. Each clinic detail page (e.g., Batignolles Paris) embeds rich MedicalBusiness + AggregateRating + Review schema with per-clinic opening hours, geo-coordinates, and service catalogs. The /soins-veterinaires/ page uses a Product schema with AggregateRating (4.6 / 100,360 reviews). However, the /maladies/ page has zero schema markup — a missed opportunity for MedicalCondition or HealthTopicContent types. The /conseils-veterinaires/ advice hub carries only a BreadcrumbList. No FAQ schema is used anywhere despite the site having FAQ-style Q&A sections on multiple pages.

Cold-Knowledge Gap

The LLM knows nothing about Sevetys. The site itself is a major French veterinary group — 250 clinics nationwide, 2,000+ staff, founded 2016, originally FamilyVets, rebranded to Sevetys in 2021. The recruitment subdomain (nous-rejoindre.sevetys.fr) contains the full origin story and scale data, but this content is siloed on a separate subdomain with no cross-linking from the main site's schema or sitemap. The main sitemap (2,209 URLs) covers clinics, care pages, advice articles, and diseases — but the corporate narrative lives entirely off the primary domain.

External Signals

Web search for "Sevetys" returns zero indexed results across multiple queries — no press, no reviews, no Reddit threads, no directory listings. The site itself has index, follow meta tags and a large sitemap, yet appears to have no organic search footprint. The DNS TXT records show an anthropic-domain-verification token, indicating Anthropic has already verified the domain for Claude crawler access — but the cold model still has no knowledge of the brand. The google-site-verification token is present, confirming Search Console setup, but Google's index appears empty or minimal.

Surprising Signals

Several advice articles carry future dates — "05/05/2026" and "06/05/2026" — suggesting either a content scheduling system leaking future publish timestamps or a date-handling bug. The site links externally to vetalia.com (a pet insurance partner) and myvetshop.fr (an online pharmacy), creating a fragmented citation graph that AI engines would struggle to consolidate. The robots.txt blocks the booking path (/prendre-rendez-vous/*) from all crawlers, meaning AI models cannot surface the core conversion flow — appointment booking — in their responses.

Findings

1. Zero LLM knowledge of Sevetys brand despite large scale High

   A frontier model queried on 'Sevetys' returns no recall — no sector, products, or reputation — despite the site representing a veterinary group with ~250 clinics, 2,000+ professionals, and over 1 million animals treated.

   What to change: Publish a comprehensive llms.txt file and ensure the corporate narrative (size, history, locations) is prominently featured on the main domain with structured data.

2. No search engine index presence for brand queries High

   Web searches for 'Sevetys' return zero indexed results across multiple queries — no press, reviews, or directory listings — despite the site having index, follow meta tags and a large sitemap.

   What to change: Investigate indexing issues (e.g., server configuration, content quality, backlink profile) and implement a link-building strategy to improve organic visibility.

3. Corporate narrative siloed on separate subdomain High

   The recruitment subdomain (nous-rejoindre.sevetys.fr) contains the full origin story and scale data, but this content is not cross-linked from the main site's schema or sitemap, fragmenting the brand story for AI crawlers.

   What to change: Add the corporate narrative to the main domain (e.g., an /about page) with Organization schema, and include the subdomain URLs in the main sitemap.

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

   The site returns a 404 for llms.txt, missing an opportunity to explicitly guide AI crawlers to key content and provide a structured brand summary.

   What to change: Create an llms.txt file that summarizes the brand, lists key pages, and provides a brief description of the veterinary group.

5. Appointment booking path blocked from all crawlers Medium

   The robots.txt disallows /prendre-rendez-vous/* for all user agents, preventing AI models from surfacing the core conversion flow in responses.

   What to change: Allow AI crawlers (e.g., GPTBot, ClaudeBot) to access the booking path while keeping it blocked for other bots if needed.

6. Disease page lacks medical schema markup Medium

   The /maladies/ page has zero schema markup, missing an opportunity to use MedicalCondition or HealthTopicContent types for better AI understanding.

   What to change: Add MedicalCondition or HealthTopicContent schema to disease pages with relevant properties.

7. FAQ schema not used despite FAQ-style content Medium

   Multiple pages contain FAQ-style Q&A sections but lack FAQPage schema, reducing the chance of rich results and AI extraction.

   What to change: Add FAQPage schema to pages with question-and-answer content.

8. Advice articles display future publication dates Low

   Several advice articles show dates like '05/05/2026' and '06/05/2026', suggesting a content scheduling bug that could confuse AI crawlers and users.

   What to change: Fix the date-handling logic to display actual publish dates or remove future timestamps.

9. External links to partners fragment citation graph Low

   The site links externally to vetalia.com and myvetshop.fr, creating a fragmented citation graph that AI engines would struggle to consolidate under the Sevetys brand.

   What to change: Ensure partner pages are clearly attributed with schema markup (e.g., isPartOf) to maintain brand association.

What's working

• All major AI crawlers receive full HTML content — All tested AI crawlers (GPTBot, ClaudeBot, etc.) receive HTTP 200 with full HTML content identical to browser baseline, with no bot-specific blocks.
• Homepage carries MedicalBusiness JSON-LD schema — The homepage includes a MedicalBusiness schema block with name, address, founding date, VAT ID, and social links, aiding entity recognition.
• Clinic detail pages embed rich schema with reviews — Each clinic page (e.g., Batignolles Paris) includes MedicalBusiness, AggregateRating, and Review schema with opening hours, geo-coordinates, and service catalogs.
• Soins vétérinaires page uses Product schema with ratings — The /soins-veterinaires/ page uses Product schema with AggregateRating (4.6 / 100,360 reviews), providing strong social proof.
• Anthropic domain verification token present — DNS TXT records include an anthropic-domain-verification token, indicating the site has been verified for Claude crawler access.
• Google Search Console verification token present — The DNS records include a google-site-verification token, confirming Search Console setup.
• Sitemap contains 2,209 URLs covering key content — The sitemap includes URLs for clinics, care pages, advice articles, and diseases, providing good coverage for crawlers.