notaires.fr — AI Site Grade

Analysis

The /fr/immobilier page — a top navigation pillar — returns HTTP 500 for every crawler and every browser, making it a dead zone for AI engines trying to understand the site's real-estate content.

Crawler Access

All major AI bots (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, anthropic-ai, Bytespider, Applebot-Extended, ChatGPT-User) receive HTTP 200 with full content on the homepage and all accessible pages. The site runs on Drupal 10 behind Incapsula CDN/WAF (visible in the 404 page's _Incapsula_Resource script and x-iinfo header). No bot is blocked, throttled, or served a JS shell — the plain HTML renders fully at ~159KB for all agents. The robots.txt uses a single User-agent: * rule with no AI-bot-specific directives; it disallows /search/, /admin/, and filtered directory query parameters but leaves all content pages open. No llms.txt exists (404).

Content & Schema Posture

The homepage carries only a BreadcrumbList schema — no Organization, WebSite, LocalBusiness, or FAQPage markup despite the page containing FAQ-style Q&A blocks. The succession (/fr/succession) and donation (/fr/donation) pages also use only BreadcrumbList plus a minimal WebPage graph. Individual notary profile pages (e.g., /fr/notaire/claude-nadin) do include a LegalService schema with name, address, and PostalAddress — this is the strongest structured data on the site. The English-language homepage (/en) shows stale news items from 2022-2023 (e.g., "Real estate market report n°59 - april 2023", "French property market report No57" from November 2022) alongside current 2026 articles, creating a mixed freshness signal for AI crawlers indexing the English section.

Cold-Knowledge Gap

The LLM's prior knowledge about Notaires de France is accurate and detailed: it correctly describes the profession's regulatory status, the fee structure (tarif reglemente), the directory tool, and the distinction from notariat.fr. However, the model knows about a "Simulateur de frais de notaire" (notary fee calculator) and a "Mon Notaire" personal account — neither of which appears prominently or is clearly findable on the homepage or main navigation. The site's actual homepage emphasizes an official directory search, apostille/legalization services, and news articles, but the interactive tools the model expects (fee calculator, personal account dashboard) are buried or linked via subdomain tools (monnotaire-masuccession.notaires.fr). The model also references "online appointment booking" as a feature, which the site's notary profile pages do not directly offer — they show phone and email but no embedded booking widget.

Critical Broken Page

The /fr/immobilier (real estate) page — a top-level navigation pillar linked from every page — returns HTTP 500 for all user-agents including browsers. This is the central hub for buying, selling, and renting property content. Its English counterpart /en/real-estate works correctly (HTTP 200, 2,362 words, rich sub-article structure). The French real-estate section is therefore invisible to AI crawlers despite being fully allowed in robots.txt. The sitemap includes 19,299+ URLs, predominantly individual notary profile pages (/fr/notaire/...), which are thin on content (~110 words each) and unlikely to rank for informational queries.

External Signals

The site has minimal external citation footprint in English-language search results. No recent reviews, Reddit threads, or press articles about notaires.fr specifically surfaced. The domain is verified with Google Search Console (google-site-verification=sinfTGT...), GlobalSign, and Sendinblue (Brevo) for email marketing. MX records route through Retarus (enterprise email security). The Wayback Machine shows the site has been live for years with regular snapshots.

Findings

1. /fr/immobilier returns HTTP 500 for all user-agents High

   The French real-estate page, a top-level navigation pillar, returns HTTP 500 for every crawler and browser, making it invisible to AI engines despite being allowed in robots.txt.

   What to change: Fix the server error on /fr/immobilier to return HTTP 200 with full content, matching the working English counterpart /en/real-estate.

2. No llms.txt file published Low

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a curated list of important pages and metadata.

   What to change: Create an llms.txt file at the root listing key pages (e.g., /fr/succession, /fr/donation, /en/real-estate) with brief descriptions.

3. English homepage shows stale news items from 2022-2023 Medium

   The English-language homepage (/en) features news articles from 2022-2023 alongside current 2026 content, creating a mixed freshness signal for AI crawlers.

   What to change: Remove or archive outdated news items from the English homepage to present a consistently current content set.

4. Homepage lacks Organization and WebSite schema Medium

   The homepage only includes BreadcrumbList schema, missing Organization, WebSite, and LocalBusiness markup that would help AI engines understand the site's identity and purpose.

   What to change: Add Organization, WebSite, and LocalBusiness structured data to the homepage with the official name, logo, and description.

5. FAQ-style content on homepage lacks FAQPage schema Medium

   The homepage contains FAQ-style Q&A blocks but does not use FAQPage structured data, reducing the chance of appearing in AI-generated answer snippets.

   What to change: Add FAQPage schema to the FAQ sections on the homepage and relevant subpages.

6. Individual notary profile pages are content-thin Medium

   Notary profile pages (e.g., /fr/notaire/claude-nadin) contain only ~110 words each, offering limited informational value for AI crawlers despite being the most numerous URLs in the sitemap.

   What to change: Enrich notary profile pages with additional details such as services offered, languages spoken, and a brief biography to improve content depth.

7. Expected interactive tools (fee calculator, personal account) not prominently linked Low

   AI models expect a notary fee calculator and personal account dashboard, but these tools are not clearly findable from the homepage or main navigation; they are buried on subdomains.

   What to change: Add prominent links to the fee calculator and Mon Notaire account on the homepage and main navigation.

8. Notary profile pages lack online appointment booking Low

   AI models expect online appointment booking, but notary profile pages only show phone and email, with no embedded booking widget.

   What to change: Integrate an online appointment booking system or link to a third-party booking service on notary profile pages.

What's working

• All major AI bots receive HTTP 200 with full HTML content — The site serves plain HTML to all tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) without blocking or throttling, ensuring content is indexable.
• Notary profile pages include LegalService schema with address — Individual notary pages use LegalService structured data with name, address, and PostalAddress, providing clear entity information for AI engines.
• English real-estate page (/en/real-estate) returns HTTP 200 with rich content — The English counterpart of the broken French page is fully accessible, containing 2,362 words and a rich sub-article structure.
• LLM prior knowledge about Notaires de France is accurate and detailed — The model correctly describes the profession's regulatory status, fee structure, directory tool, and distinction from notariat.fr, indicating strong offline brand recognition.
• robots.txt allows all content pages to be crawled — The robots.txt only disallows /search/, /admin/, and filtered directory parameters, leaving all content pages open for AI crawlers.
• Sitemap contains 80+ URLs with index, covering key pages — The sitemap is well-formed and includes a variety of pages, though dominated by notary profiles.