cockpitstudios.org — AI Site Grade

Analysis

Cockpit's AI-Visibility Audit: A Social Enterprise Confused With a VFX Studio

The most consequential finding is that frontier LLMs conflate Cockpit Studios (a London craft social enterprise founded in 1986) with a completely unrelated VFX/post-production company also called "Cockpit Studios" — a confusion that no element of the site's current AI posture works to correct.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receive a full 200 response with identical byte content to a browser baseline on the main domain. No UA-based blocking exists. The site runs on AWS CloudFront behind Cloudflare DNS, serving server-rendered HTML (WordPress) with no JS-shell risk. The robots.txt is minimal: a single Disallow: /wp-admin/ rule for * with no AI-bot-specific directives whatsoever. The shop subdomain (shop.cockpitstudios.org) is fully blocked by Cloudflare for all user-agents including browsers (403 challenge page), making product inventory invisible to crawlers.

Cold-Knowledge Gap

Asked cold, a frontier LLM describes Cockpit Studios as "a UK-based creative production company specializing in high-end VFX, CGI, and motion design" with clients including BBC, Channel 4, and work on "John Lewis Christmas Advert" and "Doctor Who." This is a different entity entirely — the actual Cockpit is a social enterprise providing affordable studio space to 175+ craft makers in Bloomsbury and Deptford. When prompted with "Cockpit Studios London craft social enterprise," the same model correctly identifies the organization (founded 1986 by the Crafts Council, independent charity, makers in ceramics/furniture/textiles/jewellery), but the default cold retrieval returns the VFX studio. The site does nothing to disambiguate itself for AI engines.

Schema Posture

The site uses Organization schema with sameAs links to Facebook, Instagram, and LinkedIn — but no @type of NGO, Nonprofit, or SocialEnterprise despite being an award-winning social enterprise and registered charity. The name field reads simply "Cockpit" with no disambiguating description like "Cockpit Studios (Craft Social Enterprise)." No Product schema exists on maker pages. No FAQPage schema is used on the FAQ page despite containing 50+ Q&A pairs in plain HTML. No Event schema appears on event pages. The WebSite schema's description reads "London's centre for excellence in craft" — accurate but insufficient for entity resolution.

Content & Structure

The homepage positions the brand clearly: "cockpit is a centre for excellence in craft," "award-winning social enterprise," "home to over 175 independent creative businesses in London," "UK's only business incubator for craft." The awards page lists 20+ named craft awards (ALMAW Award, Clothworkers' Award, Glass Sellers' Bursary, etc.) — rich, structured content that AI crawlers can access but that lacks any schema markup. The FAQ page is a dense list of plain-text Q&A with no FAQPage schema. Maker profile pages (e.g., Alison Flanders) contain detailed biographical text, craft descriptions, and contact info but carry zero JSON-LD of any type.

External Signals

The site has no detectable external press footprint in recent search results. No Reddit threads, no news articles, no review sites surfaced. The social enterprise's external mentions appear to be minimal or behind search-engine blind spots. The shop.cockpitstudios.org subdomain (likely Shopify) is entirely inaccessible to all crawlers and browsers due to Cloudflare challenge pages, meaning AI models cannot access product or pricing data from the e-commerce arm.

Findings

1. Frontier LLMs conflate Cockpit Studios with an unrelated VFX studio High

   Cold knowledge retrieval returns a description of a VFX/post-production company instead of the actual craft social enterprise. The site provides no disambiguation signals (e.g., schema, structured data) to correct this.

   What to change: Add Organization schema with @type Nonprofit or SocialEnterprise, include a disambiguating description in the name field, and publish an llms.txt file that explicitly states the organization's identity.

2. No llms.txt file published Medium

   The site does not serve an llms.txt file, missing an opportunity to provide AI crawlers with explicit guidance about the organization's identity and key pages.

   What to change: Create and publish an llms.txt file at the root that describes the organization as a craft social enterprise and lists key pages.

3. Organization schema lacks Nonprofit or SocialEnterprise type High

   The site uses Organization schema but does not specify @type Nonprofit or SocialEnterprise, despite being a registered charity and social enterprise. This reduces entity resolution accuracy for AI models.

   What to change: Update the Organization schema to include @type Nonprofit or SocialEnterprise, and add a description that disambiguates the organization from the VFX studio.

4. FAQ page lacks FAQPage schema markup Medium

   The FAQ page contains over 50 Q&A pairs in plain HTML but no FAQPage schema, reducing the likelihood of rich results and structured extraction by AI crawlers.

   What to change: Add FAQPage schema markup to the FAQ page, wrapping each Q&A pair in the appropriate JSON-LD structure.

5. Maker profile pages have no Product or Person schema Medium

   Individual maker pages (e.g., Alison Flanders) contain detailed biographical and craft information but carry zero JSON-LD markup, making it harder for AI models to extract structured data about makers and their products.

   What to change: Add Person and Product schema to maker profile pages, including name, description, craft type, and contact information.

6. Shop subdomain is fully blocked by Cloudflare challenge High

   The shop subdomain (shop.cockpitstudios.org) returns a 403 challenge page for all user-agents, including browsers and AI crawlers, making product inventory and pricing invisible to AI models.

   What to change: Allow AI crawlers (e.g., GPTBot, ClaudeBot) to access the shop subdomain by adjusting Cloudflare settings, or move product data to the main domain with proper schema.

7. Minimal external press and backlink footprint Medium

   Web searches for the organization and its domain return zero results from news, reviews, or social media, indicating low external visibility that could help AI models verify the organization's identity.

   What to change: Build external signals through press releases, partnerships, and listings on directories like Wikipedia or craft industry sites.

8. Event pages lack Event schema markup Low

   The site likely has event pages (e.g., awards, exhibitions) but no Event schema was detected, reducing the chance of rich snippets and structured extraction.

   What to change: Add Event schema to any event-related pages, including name, date, location, and description.

9. Robots.txt is minimal with no AI-bot directives Low

   The robots.txt only disallows /wp-admin/ for all user-agents and does not include any AI-bot-specific rules, which is fine for access but misses an opportunity to guide crawlers to important pages.

   What to change: Consider adding explicit allow rules for key pages (e.g., /makers/, /about-us/) and referencing the llms.txt file.

What's working

• All major AI crawlers have full access to the main domain — GPTBot, ClaudeBot, PerplexityBot, and others receive a 200 response with identical content to browsers, ensuring the site's content is fully crawlable.
• Site serves server-rendered HTML with no JavaScript shell — The WordPress site returns static HTML content, avoiding the common JS-shell problem that hides content from crawlers.
• Awards page contains detailed, structured award listings — The awards page lists 20+ named craft awards with descriptions, providing rich content that AI crawlers can access and potentially use for knowledge graph enrichment.
• Homepage clearly states the organization's mission and identity — The homepage describes Cockpit as a 'centre for excellence in craft' and 'award-winning social enterprise,' providing strong textual signals for AI models.
• XML sitemap is present and contains 80 URLs — The sitemap is accessible and lists 80 URLs, helping crawlers discover all pages on the site.
• Organization schema is present with sameAs links — The site includes Organization schema with sameAs links to social media profiles, providing basic structured data for entity recognition.