golin.com — AI Site Grade

Analysis

Golin: A PR Agency That Preaches AI Visibility but Leaves Its Own Foundation Unaddressed

Golin's own site publishes a detailed manifesto arguing that "PR is the new SEO" and that llms.txt is a critical AI-optimization tactic — yet the agency's own domain returns a 404 for /llms.txt, and its robots.txt contains zero AI-bot-specific directives.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Applebot-Extended — receive a full 200 response with identical byte content (53,558 bytes) to a browser baseline. The sole exception is Bytespider, which gets a Cloudflare 403 block. The site runs on Cloudflare (CDN/WAF) with WP Engine hosting, behind Azure DNS. No UA-based blocking exists for any mainstream AI crawler. The robots.txt uses a catch-all User-agent: * with Disallow: (empty), meaning no restrictions — but also no explicit allowances or crawl-delay for specific AI bots.

Cold-Knowledge Gap

The LLM prior knows Golin as a 1956-founded IPG agency known for "data-driven creativity," the "Golin 4D model," and clients like McDonald's, Samsung, and Kraft Heinz. It recalls the PRWeek Global Agency of the Year win and a controversial fake-influencer campaign. What the prior misses entirely: Golin now positions itself as an AI transformation consultancy with a Chief AI Officer (Jeff Beringer), a Lighthouse 360 service for human-side AI adoption, a First Answer product for AI-search visibility, and a "PR is the new SEO" thought-leadership thesis. The site's own narrative has evolved well beyond what the model cold-recalls.

Schema Posture

Every page carries Yoast-generated JSON-LD with WebPage, BreadcrumbList, WebSite, and Organization types. The Organization schema includes logo, name, and URL but is missing sameAs social profile links, foundingDate (founded 1956), numberOfEmployees (1,700+), duns, or description beyond the tagline "GO ALL IN." The work case-study pages lack CreativeWork schema. The expertise page lists 18 service categories in plain text with no Service or Offer schema markup. No FAQPage schema exists anywhere despite the site containing definitional content.

Content & Structure

The homepage is a JS-light, text-rich page (289 words visible) with a clear H1 ("Together, we create change that matters"), award badges, and a video embed from Vimeo. The Work page lists 20+ case studies with filterable categories but each case study page is thin on visible text (the Specsavers page has only 143 words) — the real campaign details live in embedded Vimeo videos that AI crawlers cannot parse. The News section is actively maintained with articles dated into 2026, covering AI, media trends, and sustainability. The Expertise page mentions "First Answer" (AI visibility optimization) and "Lighthouse 360" but these are linked as separate pages with no dedicated /first-answer URL found in the sitemap.

External Signals

The DNS TXT records reveal an anthropic-domain-verification token, confirming Golin has proactively verified its domain with Anthropic for Claude crawler access. Other verification tokens for Google, Facebook, Figma, Zoom, Adobe, Smartsheet, Miro, Jamf, and Wiz indicate a broad SaaS stack. The site's own thought-leadership article on the "AI Reputation Economy" (published June 2025, 1,138 words) explicitly recommends implementing llms.txt and avoiding JavaScript-heavy sites — recommendations the agency itself has not followed on its own domain.

Findings

1. Missing /llms.txt file High

   Golin's own site returns a 404 for /llms.txt, despite the agency's thought-leadership article recommending it as a critical AI-optimization tactic.

   What to change: Create an /llms.txt file that lists key pages and a brief description of the site's content for AI crawlers.

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

   The robots.txt uses a catch-all rule with no explicit allowances or crawl-delay for AI bots like GPTBot, ClaudeBot, or Google-Extended.

   What to change: Add explicit Allow directives for beneficial AI crawlers and consider a crawl-delay to manage load.

3. Cold-knowledge gap on AI transformation services High

   LLM prior knowledge misses Golin's AI transformation consultancy, Lighthouse 360, First Answer product, and 'PR is the new SEO' thesis, which are central to the site's current narrative.

   What to change: Ensure key AI service pages are well-linked, text-rich, and included in the sitemap; consider adding structured data for services.

4. Missing sameAs and other properties in Organization schema Medium

   The Organization JSON-LD lacks sameAs social profile links, foundingDate, numberOfEmployees, and description, reducing entity richness for AI crawlers.

   What to change: Add sameAs links to LinkedIn, Twitter, etc., and include foundingDate, numberOfEmployees, and description in the Organization schema.

5. Case study pages lack CreativeWork schema Medium

   Work case-study pages do not include CreativeWork or Article schema, missing an opportunity to signal rich content to AI crawlers.

   What to change: Add CreativeWork or Article schema to case study pages with properties like name, description, and video.

6. Expertise page lacks Service or Offer schema Medium

   The expertise page lists 18 service categories in plain text with no structured data markup, making it harder for AI to understand the service offerings.

   What to change: Add Service or Offer schema for each service category with name and description.

7. Case study pages are text-thin with video-dependent content High

   Case study pages have minimal visible text (e.g., 143 words on Specsavers page) and rely on embedded Vimeo videos that AI crawlers cannot parse.

   What to change: Add text summaries, transcripts, or captions for each case study to make the content accessible to crawlers.

8. No dedicated URL for First Answer product Medium

   The First Answer AI visibility product is mentioned on the expertise page but has no dedicated page in the sitemap, limiting its discoverability.

   What to change: Create a dedicated landing page for First Answer with detailed content and include it in the sitemap.

9. Bytespider crawler blocked by Cloudflare Low

   The Bytespider crawler receives a 403 response from Cloudflare, potentially limiting visibility on Baidu's AI ecosystem.

   What to change: If Bytespider is considered beneficial, allow it through Cloudflare WAF.

10. No FAQPage schema on definitional content Low

    Despite containing definitional content about AI and PR, the site does not use FAQPage schema, missing a chance for rich results.

    What to change: Add FAQPage schema to pages that answer common questions about AI visibility.

What's working

• All major AI crawlers allowed with full content access — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others receive a 200 response with identical content to browsers, ensuring no artificial barriers.
• Anthropic domain verification token present — DNS TXT records include an anthropic-domain-verification token, indicating proactive verification for Claude crawler access.
• Yoast-generated JSON-LD on every page — All pages include WebPage, BreadcrumbList, WebSite, and Organization schema, providing a baseline structured data foundation.
• Actively maintained news section with AI-related content — The news section is updated regularly with articles covering AI, media trends, and sustainability, providing fresh content for crawlers.
• Homepage is text-rich and JS-light — The homepage contains 289 words of visible text with a clear H1, making it easily parseable by AI crawlers.
• Published thought-leadership article on AI visibility — A 1,138-word article on the 'AI Reputation Economy' positions Golin as a knowledgeable voice and provides content for AI crawlers to index.
• Cloudflare CDN and WAF in place — The site uses Cloudflare for performance and security, which can help manage crawler traffic and mitigate abuse.
• Sitemap includes 30+ URLs with good coverage — The sitemap lists 30 URLs covering key pages, aiding crawler discovery.