dykema.com — AI Site Grade

Analysis

I have enough to write a strong audit. Let me compile the findings.

Dykema's AI crawlers get full 200 access with real content, but the site has a critical structural problem: every individual attorney profile URL redirects to the people listing page, making 415+ attorney bios invisible to AI engines as standalone entities.

Crawler Access

All 11 AI bot user-agents tested (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User) receive 200 status with identical byte-size content as a browser baseline. The site runs on Cloudflare with a RubyApps backend and a strict Content-Security-Policy. No bot is blocked, throttled, or served a JS shell. The robots.txt contains only a generic User-agent: * rule blocking four low-value paths (/print/content/, /public/docgen/, /historic-privacy-policy.html, /test-graph.html) with a Crawl-delay: 10. No AI-specific directives exist. The llms.txt returns 404.

Broken Attorney Profile Architecture

The sitemap lists 7,883 URLs, but individual attorney profile pages at paths like /people/len-wolfe or /people/richard-j-aron silently redirect to the /people/index.html listing page with a 200 status. This means no attorney bio has a canonical, crawlable URL with unique schema markup. The people listing page shows 415 results but renders only base64-encoded email addresses and minimal data -- no Person schema, no individual bios, no practice descriptions per attorney. An AI crawler cannot extract structured information about any single lawyer at the firm.

Cold-Knowledge Gap

The LLM knows Dykema as a "national law firm based in the United States, with headquarters in Chicago" strong in automotive, commercial litigation, IP, labor, real estate, tax, and healthcare. The actual site says headquarters are not stated -- the firm lists 13 offices across 8 states with no single HQ designation. The cold knowledge correctly identifies automotive as a core strength, but the site's homepage heavily emphasizes centennial messaging (100 years), which the model does not mention. The model also does not know about Dykema's cannabis practice (Law360 "Practice Group of the Year" three times in four years), its Texas Business Court leadership, or its $1.387B energy portfolio acquisition work.

Schema Posture

The homepage carries a LegalService schema type with sameAs links to Instagram, LinkedIn, and YouTube. However, the schema is thin: no address, telephone, openingHours, numberOfEmployees, areaServed, or priceRange properties. The WebPage schema on every page uses the identical generic description string. No FAQPage, Article, BlogPosting, or Person schema exists anywhere on the site. The centennial microsite at dykema100.com has a LocalBusiness schema with empty address and openingHours fields.

Content Signals

The site produces substantial content -- 1,386 press releases, 1,170 legal alerts, 764 articles, 674 blog posts, 138 "One Minute Matters" videos, 48 podcasts. The news-insights hub is well-organized with filterable content types. However, the homepage and services page rely on accordion/collapse UI for practice area listings, which may cause content visibility issues for crawlers that do not execute JavaScript. No FAQ, comparison, or table answer-format signals were detected on any page.

Findings

1. All attorney profile URLs redirect to people listing page High

   Individual attorney profile pages at paths like /people/len-wolfe silently redirect to /people/index.html with a 200 status, making each attorney's bio uncrawlable as a standalone entity. No Person schema exists on any profile.

   What to change: Remove the redirect for individual attorney profile URLs so each profile serves unique content. Add Person schema markup to each profile page.

2. No Person schema on any page High

   The site lacks Person schema markup on attorney profiles and the people listing page. AI crawlers cannot extract structured data about individual lawyers.

   What to change: Add Person schema with name, jobTitle, description, and image to each attorney profile page.

3. Homepage LegalService schema lacks key properties Medium

   The homepage's LegalService schema is missing address, telephone, openingHours, numberOfEmployees, areaServed, and priceRange. The WebPage schema uses a generic description on every page.

   What to change: Expand the LegalService schema to include address, telephone, openingHours, numberOfEmployees, and areaServed. Use unique descriptions per page in WebPage schema.

4. No FAQPage, Article, or BlogPosting schema on content pages Medium

   Despite having hundreds of articles, blog posts, and press releases, the site does not use Article or BlogPosting schema. No FAQPage schema was detected anywhere.

   What to change: Add Article or BlogPosting schema to news and insights pages. Consider FAQPage schema for practice area pages.

5. llms.txt file returns 404 Medium

   The site does not provide an llms.txt file, which is a recommended standard for AI crawlers to discover key content.

   What to change: Create an llms.txt file listing important pages like practice areas, offices, and attorney profiles.

6. Practice area listings use accordion UI that may hide content from crawlers Medium

   The homepage and services page use accordion/collapse UI for practice area listings. If content is loaded via JavaScript, some crawlers may not see it.

   What to change: Ensure practice area content is present in the initial HTML or use progressive enhancement so crawlers can access it without JavaScript.

7. Site does not state headquarters location Low

   The site lists 13 offices across 8 states but does not designate a headquarters. The LLM cold knowledge incorrectly assumes Chicago as HQ.

   What to change: Add a headquarters designation to the about or offices page, and include it in the LegalService schema.

8. Centennial microsite LocalBusiness schema has empty fields Low

   The dykema100.com site has a LocalBusiness schema with empty address and openingHours fields.

   What to change: Populate the address and openingHours fields in the LocalBusiness schema, or remove the schema if not applicable.

9. Low external visibility in AI-related searches Low

   Web searches for Dykema in AI contexts (reviews, news, trends) returned zero results, indicating limited presence in AI training data and citations.

What's working

• All 11 AI bots receive full 200 access with real content — Every tested AI crawler gets a 200 status with identical content as a browser, with no blocking or throttling.
• Robots.txt is minimal and does not block AI bots — The robots.txt only blocks four low-value paths and has no AI-specific restrictions.
• Site hosts thousands of articles, alerts, and press releases — The news-insights hub contains 1,386 press releases, 1,170 legal alerts, 764 articles, 674 blog posts, 138 videos, and 48 podcasts, providing rich content for AI training.
• Homepage includes LegalService schema with sameAs links — The homepage has a LegalService schema type with sameAs links to social media profiles, providing basic structured data.
• Sitemap is accessible and lists 80 URLs — The sitemap at /sitemap.xml returns 200 and contains 80 URLs, aiding crawler discovery.
• Site uses Cloudflare with proper security headers — Cloudflare provides CDN and security, and the site returns proper headers including Content-Security-Policy.