cplusc.com — AI Site Grade

Analysis

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

C+C: An AI-Visible Agency That AI Models Cannot See

The site has an OpenAI domain verification TXT record (openai-domain-verification=dv-N5x2CDwLQNYVWsQM7RCfwhb8) and publishes blog posts about Generative Engine Optimization and showing up in AI answers — yet a frontier LLM queried cold has zero knowledge of the brand, its clients, its campaigns, or even its existence.

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 size (285KB) to a browser baseline. Only Bytespider (TikTok/Bytedance) is blocked at Cloudflare with a 403. The robots.txt contains no AI-bot-specific rules whatsoever — just a Yoast-generated catch-all with a single Disallow for /wp-content/uploads/wpforms/. The llms.txt returns a 404. The site runs on Cloudflare (CDN + WAF) behind WP Engine hosting, with strong security headers including HSTS preload and CSP.

Cold-Knowledge Gap

The LLM queried cold reported: no awareness of any campaigns, clients, awards, or reputational signals. This is the core disconnect. The site positions itself as a behavior-change agency founded in 2005 by Julie Colehour and Bryan Cohen, with offices in Seattle, Portland, and Boston, serving clients like REI, Alaska Airlines, King County Metro, Washington State Department of Health, Massachusetts EEA, Puget Sound Energy, and Waste Management. The work page lists over 20 case studies with named clients. The team page lists ~80 staff members across multiple disciplines. None of this surfaces in the model's prior knowledge.

Schema Posture

Every page uses WebPage, BreadcrumbList, and WebSite schema with SearchAction. Blog posts use Article schema with author Person markup including job titles and images. However, no Organization or LocalBusiness schema exists anywhere — despite the site having a physical address, phone numbers, and a founding story. No FAQPage, Product, or Service schema is used on the services page, which describes eight distinct service lines in prose without structured markup.

Content & AI Signals

The blog actively covers AI visibility topics: "From Search to Strategy: How to Show Up in AI Answers" (Oct 2025) discusses GEO vs SEO, "Harnessing Generative AI" (Oct 2023) covers prompt engineering, and "How would leading Generative AI tools rewrite the Gettysburg Address?" (Jul 2025) compares GPT, Gemini, and Claude outputs. The site also offers a downloadable "Search Informed Content Framework & Checklist" and an "Equitable Engagement Framework" PDF. The /search-visibility/ landing page explicitly targets GEO optimization. Despite this content depth, the clients page (/work/clients/) contains only 6 words of visible text — "Our Past and Present Clients Include:" — with client logos likely rendered as images invisible to AI crawlers.

Findings

1. Frontier LLMs have no knowledge of C+C brand, clients, or campaigns High

   A cold query of a frontier LLM returned zero awareness of C+C, its clients (REI, Alaska Airlines, etc.), campaigns, or even its existence, despite the site having an OpenAI domain verification TXT record and publishing AI visibility content.

   What to change: Build external citations on authoritative sites, publish case studies on third-party platforms, and ensure structured data (Organization, LocalBusiness) is present on every page.

2. No Organization or LocalBusiness schema on any page High

   Despite having a physical address, phone numbers, and a founding story, the site lacks Organization and LocalBusiness schema. This deprives AI models of structured entity data needed to recognize the brand.

   What to change: Add Organization schema with name, logo, address, phone, founding date, and sameAs URLs to the homepage and contact page. Add LocalBusiness schema for each office location.

3. Clients page renders only 6 words of text; logos are invisible to crawlers High

   The /work/clients/ page contains only the text 'Our Past and Present Clients Include:' followed by client logos likely rendered as images. AI crawlers cannot extract client names from images, making the page effectively empty for knowledge extraction.

   What to change: Add visible text listing client names alongside or as alt text for logos. Consider using a text-based list with schema markup (e.g., Organization with client relationships).

4. No external web citations found for C+C brand or campaigns High

   Multiple web searches for 'C+C communications agency', 'Colehour Cohen', and site-specific queries returned zero results. The brand has no presence on third-party sites, directories, or news outlets, which is critical for AI model training data.

   What to change: Get listed on industry directories (e.g., Clutch, Agency Spotter), publish guest posts on authoritative PR/marketing sites, and encourage client mentions on their websites.

5. llms.txt file returns 404 Medium

   The site does not provide an llms.txt file, which is a recommended way to guide AI crawlers to key content. This is a missed opportunity to signal important pages to LLMs.

   What to change: Create an llms.txt file listing the most important pages (homepage, about, services, case studies, blog) with brief descriptions.

6. Services page lacks Service or FAQPage schema for eight service lines Medium

   The services page describes eight distinct service lines in prose but uses no structured markup (Service, FAQPage, or ItemList). This prevents AI models from extracting service offerings as structured data.

   What to change: Add Service schema for each service line with name, description, and provider. Consider FAQPage schema for common questions.

7. Locations page has no LocalBusiness schema for three offices Medium

   The locations page lists Seattle, Portland, and Boston offices with addresses but no LocalBusiness schema. This limits local AI visibility and entity recognition.

   What to change: Add LocalBusiness schema for each office with name, address, phone, and geo coordinates.

8. Robots.txt has no AI-bot-specific rules, missing opportunity to guide crawlers Low

   The robots.txt contains only a Yoast-generated catch-all with a single Disallow for /wp-content/uploads/wpforms/. No AI bot is explicitly allowed or disallowed, and no crawl-delay or sitemap hint is provided for AI crawlers.

   What to change: Add explicit Allow rules for GPTBot, ClaudeBot, and other AI crawlers, and consider adding a Crawl-Delay directive.

9. Bytespider (TikTok) is blocked by Cloudflare with 403 Low

   Bytespider receives a 403 response from Cloudflare, preventing TikTok's crawler from accessing the site. While this may be intentional, it limits visibility on TikTok's AI ecosystem.

   What to change: If TikTok visibility is desired, allow Bytespider by adjusting Cloudflare WAF rules.

What's working

• OpenAI domain verification TXT record present — The site has an OpenAI domain verification TXT record (openai-domain-verification=dv-N5x2CDwLQNYVWsQM7RCfwhb8), which is a prerequisite for being indexed by OpenAI's crawlers and used in GPT training.
• All major AI crawlers receive full 200 responses — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, and Applebot-Extended all receive 200 responses with identical byte size to browser baseline, ensuring content is accessible.
• Blog actively publishes AI visibility and GEO content — The blog contains posts like 'From Search to Strategy: How to Show Up in AI Answers' and 'Harnessing Generative AI', demonstrating topical authority on AI visibility.
• Blog posts use Article schema with author Person markup — Blog posts include Article schema with author Person markup including job titles and images, which helps AI models attribute content to specific authors.
• BreadcrumbList and WebSite schema with SearchAction on every page — Every page includes BreadcrumbList and WebSite schema with SearchAction, providing basic navigation and site identity to crawlers.
• Sitemap present with 55 URLs and index enabled — The sitemap is accessible and contains 55 URLs, helping crawlers discover all pages.
• Dedicated search visibility landing page targeting GEO — The /search-visibility/ page explicitly targets Generative Engine Optimization, showing strategic intent to be AI-visible.
• Strong security headers including HSTS preload and CSP — The site uses Cloudflare with HSTS preload and Content Security Policy, ensuring secure connections and protection against attacks.