kennedyslaw.com — AI Site Grade

Analysis

Kennedys Law LLP — AI-Visibility Audit

The site's Azure WAF silently blocks every major AI crawler with a 403 JS-challenge page, yet the robots.txt tells those same bots they are welcome everywhere — a contradiction that leaves AI models relying on stale, fragmented, or incorrect third-party knowledge about the firm.

Crawler Access

robots.txt at https://www.kennedyslaw.com/robots.txt is a bare-bones User-agent: * Allow: / with no AI-bot-specific rules. Despite this permissive posture, compare_bot_access on the homepage returned 403 for GPTBot, ClaudeBot, Google-Extended, OAI-SearchBot, PerplexityBot, Perplexity-User, ChatGPT-User, anthropic-ai, and Applebot-Extended. The 403 body is an "Azure WAF JS Challenge" page — a client-side JavaScript gate that no headless crawler can pass. Only Bytespider and a standard browser UA received the real 200 HTML. The site runs on Azure (IP 51.104.28.75, x-azure-ref header, x-cache header) with no Cloudflare or Akamai in the path. The /llms.txt returns 404.

Cold-Knowledge Gap

A frontier LLM queried cold about "Kennedys Law LLP" produced a profile that is partially wrong and notably thin. It stated the firm was founded in 1899 — the site's own JSON-LD schema says foundingDate: 1938. It claimed "over 40 offices worldwide" — the site lists 46 offices across 20 countries. It described Kennedys IQ as a "suite of legal technology tools" but omitted the platform's actual product names (Incident Manager, Portal Manager, Recovery Manager, Defence Lawyer, Fraud Detector, Quantum). It mentioned a "Kennedys Law app" that does not appear anywhere on the live site. It cited Legal 500 and Chambers recognition but could not name any specific ranking or practice area strength. The model's prior is a rough approximation, not an accurate firm profile.

Schema Posture

The homepage carries a single Organization schema block with correct address, contact, founding date, and social links. However, the schema is thin for a global law firm: no LegalService subtype, no PracticeArea or OpeningHoursSpecification, no NumberOfEmployees (the site says 2,900+ people). The "Who we are" page (/en/who-we-are/) returns a near-empty shell — 1 word of visible text, noindex,nofollow, no schema, no headings — yet it is the logical page for an AI crawler seeking the firm's narrative. The insights/articles use Article schema with author and date, which is correct, but the service pages (/en/our-expertise/services/insurance-and-claims/) carry zero schema markup.

External Signals

The firm operates two subdomains that are architecturally separate: kennedysiq.com (a WordPress site with rich schema, including WebSite and Organization for "Kennedys IQ Limited") and foresight.kennedyslaw.com (a gated report microsite). Neither is linked from the main site's sitemap. The main sitemap at /en/sitemap.xml contains 4,358 URLs — a comprehensive crawl surface — but the robots.txt only points to this one sitemap, omitting the IQ and Foresight subdomains entirely. Web searches for the firm returned zero results across multiple queries, suggesting low general-web citation density outside of legal directories.

Findings

1. Azure WAF silently blocks all major AI crawlers with 403 JS challenge High

   Every major AI crawler (GPTBot, ClaudeBot, Google-Extended, OAI-SearchBot, PerplexityBot, etc.) receives a 403 JS-challenge page from the Azure WAF, while robots.txt permits all bots. This contradiction renders the site invisible to AI crawlers.

   What to change: Remove the JS challenge for AI crawler user agents, or configure the WAF to allow known AI bot IP ranges. Alternatively, serve a static HTML version to blocked bots.

2. No /llms.txt file published Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a curated set of important URLs and context.

   What to change: Publish an /llms.txt file listing key pages such as the homepage, about page, services, and insights.

3. LLM cold knowledge is partially wrong and notably thin High

   A frontier LLM queried about Kennedys Law produced a profile with incorrect founding date (1899 vs actual 1938), fabricated product names (Kennedys Law app), and omitted key details like specific office count and IQ platform products.

   What to change: Ensure AI crawlers can access the site to refresh knowledge, and consider publishing structured data and an /llms.txt to guide LLM ingestion.

4. Organization schema lacks LegalService subtype and key properties Medium

   The homepage Organization schema is correct but minimal: no LegalService subtype, no PracticeArea, no OpeningHoursSpecification, no NumberOfEmployees. This limits the depth of AI understanding.

   What to change: Add LegalService subtype, PracticeArea, OpeningHoursSpecification, and NumberOfEmployees to the Organization schema.

5. Who we are page is a near-empty shell with noindex High

   The /en/who-we-are/ page returns only 1 word of visible text, has noindex,nofollow, no schema, and no headings. It is the logical page for AI crawlers seeking firm narrative but provides no content.

   What to change: Populate the page with substantive firm history, values, and leadership information; remove noindex; add Organization schema.

6. Service pages lack any schema markup Medium

   Pages like /en/our-expertise/services/insurance-and-claims/ have zero schema markup, missing the chance to define practice areas and services for AI.

   What to change: Add LegalService or Service schema to each service page with appropriate properties.

7. Kennedys IQ and Foresight subdomains omitted from main sitemap Medium

   The main sitemap at /en/sitemap.xml does not include kennedysiq.com or foresight.kennedyslaw.com, reducing their discoverability by crawlers.

   What to change: Include subdomain URLs in the main sitemap or create separate sitemaps referenced in robots.txt.

8. Near-zero web search results for the firm Medium

   Multiple web searches for Kennedys Law returned zero results, indicating low general-web citation density outside of legal directories.

   What to change: Increase external citations through press releases, guest articles, and directory listings to improve AI knowledge sources.

9. Robots.txt is bare-bones with no AI bot rules Low

   The robots.txt only has a single Allow: / rule for all user agents, missing opportunities to guide AI crawlers to specific sections or sitemaps.

   What to change: Add specific rules for AI bots, reference the sitemap, and consider disallowing low-value paths.

What's working

• Homepage Organization schema is correct and complete — The homepage carries a valid Organization schema with correct address, contact, founding date, and social links, providing a solid foundation for AI understanding.
• Insights articles use correct Article schema — Articles on the insights page include Article schema with author and date, which helps AI correctly attribute content.
• Sitemap contains 4,358 URLs covering the main site — The sitemap at /en/sitemap.xml is comprehensive, providing a full crawl surface for search engines that can access the site.
• Kennedys IQ subdomain has rich schema markup — The kennedysiq.com subdomain includes WebSite and Organization schema for Kennedys IQ Limited, providing structured data for that brand.
• Global offices page has substantial content — The /en/where-we-are/global-offices/ page contains 2,058 words of detailed office information, useful for AI crawlers that can access it.