fasken.com — AI Site Grade

Analysis

Fasken.com is entirely invisible to every AI crawler — the live site returns a Cloudflare JS challenge (HTTP 403) to all user-agents, including GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and even standard browsers without JavaScript execution.

Crawler Access

The entire domain sits behind Cloudflare's JS challenge (Under Attack mode or equivalent). compare_bot_access on the homepage returned 403 for all 11 tested user-agents — every AI crawler and a standard browser alike. The response body is a <title>Just a moment...</title> Cloudflare challenge page with a noindex,nofollow robots meta tag. The robots.txt at /robots.txt is accessible (HTTP 200) but contains no AI-bot-specific rules — only a catch-all User-agent: * with standard disallows for /sitecore/, /assets/, and payment paths. No mention of GPTBot, ClaudeBot, PerplexityBot, Google-Extended, or any other AI crawler. The llms.txt at /llms.txt returns the same Cloudflare 403 challenge — effectively non-existent. The sitemap at /sitemap.xml is also blocked (403), making it impossible for crawlers to discover the site's URL structure.

Cold-Knowledge Gap

The LLM knows Fasken as "one of Canada's Seven Sisters" with ~925 lawyers, offices in 9 cities across Canada, London, and Johannesburg, and a strong mining/energy practice. This prior knowledge is richer than what the live site delivers to any crawler — which is literally nothing. The model correctly identifies the firm's full legal name (Fasken Martineau DuMoulin LLP), its 2003 merger origin, and its Chambers/ Legal 500 rankings. However, the model's knowledge is stale on leadership: it names Peter Feldberg as managing partner, while the archived site shows Martin K. Denyes holds that role. The model also does not know about the firm's 2017 rebrand from "Fasken Martineau" to simply "Fasken."

Schema Posture

The archived homepage (September 2024 snapshot) contains rich JSON-LD: LegalService, Organization, WebSite, and WebPage types with full contact details for all 10 offices (including Tsuut'ina), social profiles (Facebook, LinkedIn, YouTube), and a SearchAction potential action. The practices page adds BreadcrumbList. This schema is well-structured and comprehensive — but it only exists in the Wayback Machine. The live site serves zero schema to any non-JS client because the Cloudflare wall intercepts every request before the application layer.

External Signals

Wikipedia describes Fasken as an international business law firm with ~925 lawyers, tracing its origins to 1863. The Wikipedia article carries a conflict-of-interest warning from January 2019 ("A major contributor to this article appears to have a close connection with its subject"), which may affect how AI models weight that source. Chambers and Partners and Legal 500 profile pages for Fasken return 404s — the URLs have changed or been decommissioned. The DNS TXT records reveal a heavy vendor stack: Sitecore (CMS), Microsoft 365 (Outlook mail), Adobe, Atlassian, DocuSign, Duo, OneTrust, and Canva — indicating a complex enterprise tech ecosystem that may contribute to the aggressive Cloudflare configuration.

Findings

1. Cloudflare JS challenge blocks all AI crawlers High

   The live site returns HTTP 403 with a Cloudflare JS challenge page to all 11 tested user-agents, including GPTBot, ClaudeBot, PerplexityBot, and Google-Extended. No crawler can access any page content.

   What to change: Disable the JS challenge for known AI crawler user-agents (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) by configuring Cloudflare WAF to allow them, or serve a static HTML version to those bots.

2. No AI-bot-specific rules in robots.txt High

   The robots.txt file contains only a catch-all User-agent: * with standard disallows and no mention of GPTBot, ClaudeBot, PerplexityBot, or Google-Extended. AI crawlers are not explicitly allowed or blocked, but the Cloudflare challenge makes access impossible anyway.

   What to change: Add explicit Allow rules for AI crawlers (e.g., User-agent: GPTBot Allow: /) and ensure the Cloudflare configuration respects them.

3. llms.txt returns Cloudflare 403 challenge High

   The llms.txt file at /llms.txt is blocked by the same Cloudflare JS challenge, returning HTTP 403. AI crawlers cannot discover any structured guidance for LLM consumption.

   What to change: Serve llms.txt without JS challenge by excluding it from Cloudflare WAF rules, and populate it with a summary of the firm and links to key pages.

4. Sitemap blocked by Cloudflare High

   The sitemap at /sitemap.xml returns HTTP 403, preventing crawlers from discovering the site's URL structure. This severely limits indexing even if access were granted.

   What to change: Allow access to /sitemap.xml for all crawlers by excluding it from Cloudflare JS challenge rules.

5. No structured data served to crawlers on live site High

   The live site serves zero JSON-LD or other structured data to any non-JS client because the Cloudflare wall intercepts all requests. Archived snapshots show rich schema (LegalService, Organization, WebSite) but this is invisible to AI crawlers.

   What to change: Ensure JSON-LD schema is served in the initial HTML response for all pages, and that the Cloudflare configuration allows crawlers to receive that HTML.

6. LLM knowledge of leadership is outdated Medium

   The LLM's prior knowledge names Peter Feldberg as managing partner, while the archived site shows Martin K. Denyes holds that role. This discrepancy can lead to fabricated or incorrect information in AI-generated summaries.

   What to change: Publish a current leadership page with clear structured data (Person schema) and ensure it is accessible to crawlers.

7. LLM unaware of 2017 rebrand to 'Fasken' Medium

   The LLM does not know about the firm's 2017 rebrand from 'Fasken Martineau' to simply 'Fasken'. This could cause AI outputs to use the old name, confusing users.

   What to change: Prominently display the current brand name on the site and include it in structured data (Organization name).

8. Chambers and Legal500 profile pages return 404 Medium

   The Chambers and Legal500 profile URLs for Fasken return 404 errors, indicating broken external links. This reduces the firm's visibility on authoritative legal directories and may affect AI citation of rankings.

   What to change: Update or redirect the Chambers and Legal500 profile URLs to the correct pages, and ensure the firm's profiles on those sites are active.

9. Wikipedia article has conflict-of-interest warning Low

   The Wikipedia article for Fasken carries a conflict-of-interest warning from January 2019, which may cause AI models to discount or flag the source as less reliable.

   What to change: Engage with Wikipedia editors to resolve the conflict-of-interest concerns and improve the article's neutrality.

10. No Google search results for fasken.com High

    Web searches for site:fasken.com and various queries about the firm returned zero results, indicating the site is not indexed by Google. This is consistent with the Cloudflare block.

    What to change: Resolve the Cloudflare block to allow Googlebot to crawl and index the site.

What's working

• Archived pages contain comprehensive JSON-LD schema — The September 2024 archived homepage includes well-structured JSON-LD for LegalService, Organization, WebSite, and WebPage types, with full office contact details and social profiles. This schema, if served live, would strongly support AI understanding.
• Practices page includes BreadcrumbList schema — The archived practices page includes BreadcrumbList structured data, aiding navigation understanding for crawlers.
• LLM has rich prior knowledge about the firm — The LLM correctly identifies Fasken as a leading international law firm with ~925 lawyers, offices in 9 cities, and strong mining/energy practice. This provides a foundation for AI visibility if the site becomes accessible.
• Robots.txt is accessible and not blocked — The robots.txt file returns HTTP 200 and is readable, allowing crawlers to at least see the disallow rules. This is a basic but positive signal.
• Wikipedia article provides authoritative overview — The Wikipedia article offers a detailed history and description of the firm, which AI models can use as a reliable source (despite the conflict-of-interest warning).