irwinmitchell.com — AI Site Grade

Analysis

Irwin Mitchell — AI-Visibility Audit

The site's most consequential AI-visibility gap is not a block — it is the absence of any structured AI onboarding layer (no llms.txt, no AI-specific robots.txt rules, no LegalService or Attorney schema on service pages) despite being one of the UK's largest law firms with 5,400+ indexed pages and full, unrestricted access for every major AI crawler.

Crawler Access

All eleven tested AI bots — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User — receive a 200 response with the same 804KB HTML payload as a browser. The site runs on Vercel (Next.js) behind Akamai DNS. No UA-based blocking, no Cloudflare challenge, no JS-gating. The robots.txt (User-agent: *) disallows only a handful of filtered paths (/newsandmedia/news-list, gated-content PDFs) and contains zero AI-bot directives. /llms.txt returns a 404 (rendered as a Next.js 404 page). This is a fully open site that AI crawlers can ingest at will — but with no guidance on what to prioritise.

Schema Posture

The homepage carries a Corporation + LegalService schema with an AggregateRating (4.8/5 from 12,774 reviews) and a ContactPoint. The medical negligence page is the only deep page with rich schema: a FAQPage (4 questions with answers) and a VideoObject. Every other key page — personal injury, business services, about, careers, news — carries only a single ImageObject schema. No Attorney, LegalService, Service, Article, or BreadcrumbList schemas appear on those pages. A site with 5,400+ URLs and deep practice-area content is leaving almost all of it schema-bare for AI engines.

Cold-Knowledge Gap

The LLM prior knows Irwin Mitchell as a UK claimant-focused personal injury and medical negligence firm, founded 1912 in Sheffield, with a notable SRA fine for conveyancing client-money mishandling. The site itself positions as a full-service national law firm — business services (corporate, commercial, tax, real estate, employment), family law, conveyancing, court of protection, and a King's Award for International Trade. The cold model does not know about the business/commercial practice, the 21-office national footprint, the 3,000+ staff, the King's Award, the thought-leadership reports (Leading Litigator, Inheritance Tax Revolution), or the podcast/newsletter operation. The AI knowledge is stuck on the claimant-PI reputation and a regulatory fine; the site's broader positioning as a business-law firm is invisible.

External Signals

The homepage links to a Trustpilot profile (aggregate rating embedded in schema), but Trustpilot blocks automated fetches with a JS challenge. The site references Legal 500 and Chambers rankings ("number one law firm for personal injury / medical negligence") but does not link to those rankings. Social profiles (X, Facebook, Instagram, LinkedIn, YouTube) are present in schema sameAs. No press coverage or third-party articles were surfaced in search that contradict or amplify the site's claims. The SRA fine mentioned in the cold model's prior is not addressed anywhere on the site — no mention, no rebuttal, no compliance page.

Surprising Details

The sitemap contains 5,428 URLs — a very large corpus for a law firm — but the news-and-insights section appears to be a thin listing page with no individual article URLs exposed in the sitemap sample. The careers page links to a separate subdomain (careers.irwinmitchell.com) that was not crawled. The awards page lists accolades going back to 2012 with no structured schema (Award type absent). The site uses Sitecore as a CMS (visible in TXT verification records and URL patterns) but is served through Vercel/Next.js — a dual-infrastructure setup that may complicate content freshness signals.

Findings

1. No llms.txt file for AI crawler guidance High

   The site returns a 404 for /llms.txt, providing no structured onboarding for AI crawlers to discover key content.

   What to change: Create an llms.txt file that lists the most important pages (practice areas, about, news) and provides a summary of the firm's full-service capabilities.

2. No AI-specific directives in robots.txt Medium

   The robots.txt file contains no rules for AI crawlers like GPTBot or ClaudeBot, leaving them to crawl without prioritization guidance.

   What to change: Add AI-specific user-agent directives to robots.txt to prioritize key pages and disallow low-value paths.

3. Missing LegalService and Attorney schema on practice-area pages High

   Only the homepage has LegalService schema; deep pages like personal injury, business services, and about lack Attorney, LegalService, or Service schema, reducing their visibility in AI knowledge graphs.

   What to change: Add LegalService, Attorney, and Service schema to all practice-area and service pages, including office locations and practice descriptions.

4. LLM knowledge lacks full-service business practice High

   The cold LLM prior only knows Irwin Mitchell as a claimant-PI firm with a regulatory fine, missing its business services, 21 offices, 3,000+ staff, and King's Award.

   What to change: Publish structured data and content that explicitly describes the full-service business practice, office network, and awards to correct the LLM knowledge gap.

5. News and insights pages lack Article schema Medium

   The news listing page and in-focus page have no Article or NewsArticle schema, making it harder for AI to treat them as authoritative content.

   What to change: Add Article or NewsArticle schema to all news and insight pages, including author, date, and headline.

6. Awards page lacks Award structured data Medium

   The awards page lists accolades since 2012 but uses no Award schema, reducing their discoverability in AI knowledge bases.

   What to change: Add Award schema to each award entry, including name, awarding body, date, and description.

7. No BreadcrumbList schema on any page Low

   No page tested includes BreadcrumbList schema, which helps AI understand site hierarchy and context.

   What to change: Add BreadcrumbList schema to all pages to improve navigational context for AI crawlers.

8. Trustpilot profile blocks automated access Medium

   The Trustpilot profile linked from the homepage returns a 403 with a JS challenge, preventing AI crawlers from verifying the aggregate rating.

   What to change: Ensure the Trustpilot profile is accessible to crawlers or embed the review data directly on the site with structured data.

9. SRA fine not mentioned or rebutted on site Medium

   The cold LLM prior includes a notable SRA fine for conveyancing client-money mishandling, but the site contains no mention or rebuttal of this incident.

   What to change: Publish a compliance or regulatory page that addresses the fine and outlines corrective actions taken.

10. News section appears as thin listing without individual article URLs in sitemap Medium

    The sitemap sample shows 5,428 URLs but the news-and-insights section seems to be a listing page with no individual article URLs exposed, limiting AI access to detailed content.

    What to change: Ensure individual article pages are included in the sitemap and have proper Article schema.

11. Careers subdomain not included in audit crawl Low

    The careers page links to careers.irwinmitchell.com, which was not crawled; this subdomain may contain important employer branding content invisible to AI.

    What to change: Ensure the careers subdomain is included in the sitemap and has consistent schema markup.

What's working

• All major AI crawlers have unrestricted access — All eleven tested AI bots receive a 200 response with full HTML content, with no UA-based blocking or JS challenges.
• Homepage has LegalService schema with aggregate rating — The homepage includes Corporation and LegalService schema with an AggregateRating of 4.8/5 from 12,774 reviews, providing strong social proof to AI.
• Medical negligence page has FAQPage and VideoObject schema — The medical negligence page includes a FAQPage with 4 questions and a VideoObject, enhancing its visibility in AI-generated answers.
• Sitemap contains 5,428 URLs, indicating extensive content — The sitemap lists over 5,400 URLs, suggesting a large corpus of practice-area and insight content available for crawling.
• Social media profiles included in schema sameAs — The homepage schema includes sameAs links to X, Facebook, Instagram, LinkedIn, and YouTube, helping AI connect the firm's online presence.
• Site hosted on Vercel with fast response times — The site runs on Vercel (Next.js) behind Akamai, delivering fast page loads and good uptime for crawlers.