atlanticaccess.co.uk — AI Site Grade

Analysis

Atlantic Access — AI-Visibility Audit

The most consequential finding is a total cold-knowledge identity mismatch: the leading frontier LLM describes Atlantic Access as a video-game distributor for indie titles on Steam and Nintendo eShop, while the actual site is a London-based e-commerce and digital distribution agency that launches US consumer brands into Europe. The model's prior has zero overlap with the site's real positioning.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Applebot-Extended — receive a full 200 response with identical HTML content to a browser baseline. Only Bytespider (ByteDance) returns a 403. The site runs on Apache on a single IP (92.205.168.207) with no CDN, no WAF, and no security headers (no HSTS, no CSP, no X-Frame-Options). The robots.txt is a bare 29 bytes: User-agent: * with a crawl-delay of 2 seconds and no AI-bot directives whatsoever. No llms.txt exists (404). The sitemap is a WordPress-native wp-sitemap.xml index with 8 sub-sitemaps covering 12 total URLs — a very small footprint.

Cold-Knowledge Gap

Asked cold, the LLM describes Atlantic Access as "a UK-based digital distribution agency specializing in helping independent video game developers" distributing titles like *Hollow Knight* and *Moonlighter* on Steam and Nintendo eShop. This describes a completely different company — likely a separate entity with the same name in the gaming space. The actual Atlantic Access (atlanticaccess.co.uk) is a turnkey European launch agency for US consumer brands, offering Amazon management across 13 countries, localized D2C websites, B2B sales, VAT/tax handling, and logistics. The site claims 200,000+ units shipped, 10 VAT registrations, 53 Amazon EU FBA warehouses, and a 5-year average partnership length. The model knows none of this.

Schema Posture

Zero structured data exists anywhere on the site. No Organization, LocalBusiness, Service, FAQPage, Product, or Review schema was found on any of the 12 pages crawled — not even on the homepage, about page, services pages, or brand case studies. The homepage has no meta description, no Open Graph tags, and no JSON-LD of any type. The site title is simply "Atlantic" — no descriptor, no brand differentiation in the <title> tag. This is a complete schema vacuum that leaves AI engines to infer the company's identity from weak signals.

External Signals

The site has near-zero external discoverability. Searches for the brand name, founder name (Ben Hofman), client names (Rocketbook, Skinnydip, Bitfinder, HeroClip, Mighty Audio), and the domain itself returned no indexed results on DuckDuckGo, no press mentions, no reviews, no Reddit threads, and no third-party case studies. The only external links found on the site are to a LinkedIn company page and the web developer's portfolio. The hello-world WordPress default post exists (empty body, 200 status) — a stale artifact that should not be in the sitemap. The site's PHP version (7.3.33) is outdated and end-of-life.

Findings

1. LLM cold knowledge describes Atlantic Access as a video-game distributor, not the actual e-commerce agency High

   Leading frontier LLMs describe Atlantic Access as a UK-based digital distribution agency for indie video games on Steam and Nintendo eShop, while the actual site is a London-based turnkey European launch agency for US consumer brands. The model's prior has zero overlap with the site's real positioning.

   What to change: Add comprehensive Organization and Service schema to the homepage and about page, and publish an llms.txt file that explicitly describes the company's actual services and brand portfolio.

2. No structured data of any kind on any page High

   Zero Organization, LocalBusiness, Service, FAQPage, Product, or Review schema was found on any of the 12 pages crawled. The homepage has no meta description, no Open Graph tags, and no JSON-LD. The site title is simply 'Atlantic' with no descriptor.

   What to change: Add JSON-LD Organization schema to the homepage with name, description, logo, URL, sameAs links, and founding date. Add Service schema to each service page and Brand schema to brand pages.

3. Near-zero external discoverability across search and social High

   Searches for the brand name, founder name (Ben Hofman), client names (Rocketbook, Skinnydip, Bitfinder, HeroClip, Mighty Audio), and the domain itself returned no indexed results on DuckDuckGo, no press mentions, no reviews, no Reddit threads, and no third-party case studies.

   What to change: Build a PR and content marketing strategy to generate press mentions, guest posts, and case studies. Ensure all client brands are linked from the site and that the LinkedIn page is active and linked back.

4. Robots.txt is bare with no AI-bot directives Medium

   The robots.txt is only 29 bytes, containing a single User-agent: * rule with a crawl-delay of 2 seconds. No AI-bot-specific directives exist for GPTBot, ClaudeBot, or others.

   What to change: Add explicit allow/disallow rules for AI crawlers and consider adding a crawl-delay for non-AI bots.

5. No llms.txt file published Medium

   The llms.txt file returns a 404, meaning AI engines have no explicit guidance on how to use the site's content.

   What to change: Create an llms.txt file that describes the company, lists key pages, and provides a summary for AI consumption.

6. Stale WordPress default 'Hello World' page in sitemap Low

   The /hello-world/ page exists with an empty body and a 200 status, and is included in the sitemap. This is a stale artifact that should not be indexed.

   What to change: Remove the hello-world page and exclude it from the sitemap.

7. Outdated PHP version (7.3.33) in use Medium

   The server reports PHP version 7.3.33, which is end-of-life and no longer receives security updates.

   What to change: Upgrade PHP to a supported version (8.0 or later) to ensure security and performance.

8. Missing security headers (HSTS, CSP, X-Frame-Options) Low

   The site lacks HSTS, Content-Security-Policy, and X-Frame-Options headers, which are standard for modern web security.

   What to change: Add HSTS, CSP, and X-Frame-Options headers to improve security posture.

9. Sitemap contains only 12 URLs Medium

   The WordPress sitemap index has 8 sub-sitemaps covering only 12 total URLs, indicating a very small site footprint that limits discoverability.

   What to change: Expand the site with more content pages, case studies, and service descriptions to increase the number of indexed URLs.

10. No CDN or WAF in place Low

    The site runs on a single IP with Apache and no CDN or Web Application Firewall, which can affect performance and security.

    What to change: Consider using a CDN and WAF to improve load times and security.

What's working

• All major AI crawlers receive full access — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, and Applebot-Extended all receive a 200 response with identical HTML content to a browser baseline.
• Consistent HTML content served to all bots — All tested AI crawlers receive the same HTML content as a regular browser, with no cloaking or differential serving.
• Sitemap is present and valid — A WordPress-native sitemap exists and returns 200, listing 12 URLs across 8 sub-sitemaps.