thegraygency.com — AI Site Grade

Analysis

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

The Graygency is invisible to AI models despite having technically open crawler access — a complete cold-knowledge void that contradicts the site's own positioning as a "leading" London paid-media agency.

Crawler Access

All major AI bots — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receive a 200 with full HTML content identical to browser baseline (272,895 bytes). No UA-based blocking, no Cloudflare challenge, no JS shell. The robots.txt is a Yoast-generated template containing a placeholder sitemap URL (https://www.example.com/sitemap.xml) and zero AI-bot-specific directives. The /llms.txt returns a 404 (WordPress 404 page with full theme CSS/JS). The actual sitemap at /sitemap_index.xml is well-structured with 9 sub-sitemaps covering ~200+ URLs including posts, case studies, and services.

Cold-Knowledge Gap

A frontier LLM queried cold about "The Graygency paid media agency" returned: "I don't have specific, verified information... It is not a well-known or widely documented entity in my training data." The model could not confirm services, clients, or location. This is the most consequential finding: the site describes itself as a "leading paid social advertising agency in London" with named clients (Lick, Edit Suits, Form Nutrition, Astley Clarke, Scandiborn) and specific results (66% new customer acquisition, 8.6x ROAS), yet no AI training corpus has ingested this brand identity. The gap between the site's self-presentation and what models know cold is total.

Schema Posture

JSON-LD is present via Yoast on every page but structurally thin. The Organization schema includes name, logo, URL, and sameAs — but the Instagram sameAs URL contains a typo: http://www.instagram.com/thegraygencysome (extra "some" appended). The actual Instagram handle is thegraygency. The WebSite schema describes the site as a "Facebook Advertising Agency" — a description that contradicts the homepage H1 ("Where brand & performance meet commercially focused paid-media") and the service pages which emphasize Meta, TikTok, Snapchat, and Pinterest. No Service, FAQPage, Product, or LocalBusiness schema is used on service pages despite FAQ sections and detailed service descriptions being present.

External Signals

Web searches for "The Graygency", "thegraygency.com reviews", and "thegraygency paid media agency" returned zero indexed results across multiple queries. No third-party reviews, no Clutch/GoodFirms listings, no Reddit threads, no press mentions, no backlink profile visible via search. The site's external link sample points to ecommentarypodcast.com (a podcast site), LinkedIn, Instagram, Facebook, and TikTok — but the Instagram URL in the schema is broken. The DNS resolves to Cloudflare (104.17.15.101) with Hetzner nameservers and Google Workspace mail. The site runs on WordPress with Yoast SEO Premium and uses FlyingPress caching.

Findings

1. Complete cold-knowledge void: AI models have no information about The Graygency High

   A frontier LLM queried cold about 'The Graygency paid media agency' returned no specific information, unable to confirm services, clients, or location. Despite the site describing itself as a leading London paid-media agency with named clients and results, no AI training corpus has ingested this brand identity.

   What to change: Increase off-site signals: publish guest posts, earn backlinks, get listed on agency directories (Clutch, GoodFirms), and ensure consistent NAP citations across the web.

2. Broken Instagram URL in Organization schema due to typo High

   The Organization JSON-LD includes a sameAs URL with a typo: 'http://www.instagram.com/thegraygencysome' (extra 'some' appended). The correct handle is 'thegraygency'.

   What to change: Correct the Instagram URL in the Organization schema to 'https://www.instagram.com/thegraygency/'.

3. WebSite schema describes site as 'Facebook Advertising Agency' contradicting broader positioning Medium

   The WebSite schema's description labels the site as a 'Facebook Advertising Agency', while the homepage and service pages emphasize a broader paid-media offering including Meta, TikTok, Snapchat, and Pinterest.

   What to change: Update the WebSite schema description to accurately reflect the agency's full service range, e.g., 'Paid media agency specializing in Meta, TikTok, Snapchat, and Pinterest advertising.'

4. No Service or FAQPage schema on service pages despite FAQ sections Medium

   Service pages like /service/paid-social/ contain detailed FAQ sections but lack FAQPage schema. No Service schema is used to describe the paid social offering.

   What to change: Add FAQPage schema to FAQ sections and Service schema to each service page with name, description, and provider.

5. Zero indexed external signals: no reviews, directory listings, or press mentions High

   Web searches for 'The Graygency', 'thegraygency.com reviews', and 'thegraygency paid media agency' returned zero results. No third-party reviews, Clutch/GoodFirms listings, Reddit threads, or press mentions were found.

   What to change: Claim profiles on agency directories (Clutch, GoodFirms, DesignRush), encourage client reviews, and pursue PR or guest posting to build off-site signals.

6. Missing /llms.txt file returns 404 Medium

   The /llms.txt endpoint returns a 404 WordPress page, missing an opportunity to provide AI models with structured brand information.

   What to change: Create an /llms.txt file with a summary of the agency, key services, client list, and links to important pages.

7. Robots.txt contains placeholder sitemap URL pointing to example.com Low

   The robots.txt file includes a sitemap directive pointing to 'https://www.example.com/sitemap.xml' instead of the actual sitemap.

   What to change: Update the Sitemap directive in robots.txt to point to 'https://thegraygency.com/sitemap_index.xml'.

What's working

• All major AI bots receive full HTML content without blocking — GPTBot, ClaudeBot, PerplexityBot, and other AI crawlers receive a 200 response with full HTML content, identical to browser baseline. No UA-based blocking or Cloudflare challenges are applied.
• Sitemap is well-structured with 9 sub-sitemaps covering ~200+ URLs — The sitemap at /sitemap_index.xml is properly formatted and includes sub-sitemaps for posts, case studies, services, and other content, aiding discovery.
• JSON-LD schema is present on every page via Yoast SEO — Yoast SEO Premium injects Organization and WebSite schema on all pages, providing basic structured data for search engines and AI models.
• Service pages contain detailed, long-form content with FAQs — Pages like /service/paid-social/ include over 2000 words of descriptive content and FAQ sections, providing rich information for AI models to potentially extract.
• Case studies include specific client results and metrics — Case studies like Lick - Paid Social & Creative Testing provide concrete metrics (66% new customer acquisition, 8.6x ROAS), which are valuable for AI training.