jay-jays.com — AI Site Grade

Analysis

Cold-knowledge gap is the headline finding

The LLM cold-knowledge model (DeepSeek) describes JayJays as a manufacturer of the Osprey MK4 webbing system for the British Army, located in Bideford, Devon, operating since the 1980s. The actual site reveals a fundamentally different reality: JayJays is a retailer (not primarily a manufacturer), located in Brecon, Wales (not Bideford), selling Gen 4 & 5 webbing, bergans, clothing, and outdoor gear — and the Osprey MK4 is not mentioned anywhere on the homepage or product pages. The model's prior is substantially inaccurate on location, product scope, and brand identity.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended, anthropic-ai — receive a 200 with full content identical to browser baseline (131,571 bytes). No UA-based blocking, no WAF challenge, no JS shell. The robots.txt has a single User-agent: * rule disallowing only backend paths (/bin/, /client_scripts/, /flash/, etc.) and contains zero AI-bot-specific directives. No llms.txt exists (404). The sitemap at /sitemap.xml is an empty XML envelope (161 bytes, zero URLs) — generated by Screaming Frog but containing no actual page references, meaning search engines and AI crawlers have no structured discovery path to the site's product pages.

Content & Schema Posture

The homepage presents a dense category navigation tree but no narrative brand story, no "About Us" page (404), no FAQ page (404), no blog (404), and no delivery/returns pages (both 404). The only informational page that resolves is the Click & Collect page. JSON-LD schema is present as Organization and Store types with correct address and phone, but no Product schema on any product listing or detail page — a critical gap for an ecommerce site selling hundreds of SKUs. The homepage has no H1 tag (only H2 and H3), which is an unusual structural omission. The site runs on Citrus-Lime Hosting (an ecommerce platform), hosted at 178.238.142.219 with Microsoft 365 mail.

Cold-Knowledge Gap

The LLM's prior knowledge contains three significant factual errors: (1) location stated as Bideford, Devon — the site shows Brecon, Wales (LD3 8LA); (2) the model claims JayJays manufactures the Osprey MK4 system — the site sells Gen 4 & 5 webbing and makes no mention of Osprey MK4; (3) the model says the company has been in business since the 1980s — the site provides no founding date at all. The model also describes JayJays as a manufacturer when the site positions itself as a supplier/retailer ("UK supplier of Specialist Military Webbing Equipment"). The model correctly identifies the brand's association with British Armed Forces load carriage but overstates the manufacturing role.

External Signals

External search results returned zero indexed mentions of JayJays across Reddit, review sites, or general web search. The brand has Facebook, Instagram, and YouTube channels linked in the footer, but these social profiles generated no discoverable external citations. The absence of any third-party reviews, forum discussions, or press coverage creates a near-total external signal vacuum — AI engines have almost no corroborating off-domain content to draw from, which explains the model's reliance on stale or incorrect training data.

Findings

1. Sitemap contains zero URLs High

   The sitemap at /sitemap.xml is an empty XML envelope with no URLs, preventing search engines and AI crawlers from discovering product pages.

   What to change: Generate a complete sitemap listing all product, category, and informational pages, and submit it to search engines.

2. No Product schema on any page High

   JSON-LD schema is limited to Organization and Store types; no Product schema exists on product listing or detail pages, reducing AI understanding of inventory.

   What to change: Add Product schema with name, description, price, and availability to all product pages.

3. LLM prior knowledge contains major factual errors High

   DeepSeek describes JayJays as a manufacturer in Bideford, Devon, producing the Osprey MK4 system since the 1980s, but the site shows a retailer in Brecon, Wales, selling Gen 4 & 5 webbing with no mention of Osprey MK4 or founding date.

   What to change: Publish an About Us page with accurate company history, location, and product scope, and add structured data to correct LLM misconceptions.

4. Key informational pages return 404 High

   The About Us, FAQ, blog, delivery, and returns pages all return 404 errors, leaving AI crawlers without authoritative content about the company.

   What to change: Create and publish these pages with accurate, detailed content to provide AI crawlers with authoritative information.

5. Zero external citations or reviews found High

   Web searches for JayJays on Reddit, review sites, and general web returned no results, creating a vacuum of off-domain corroboration for AI models.

   What to change: Encourage customer reviews on third-party platforms and engage in forums to build external signals.

6. No llms.txt file available Medium

   The site does not provide an llms.txt file, missing an opportunity to guide AI crawlers to key content.

   What to change: Create an llms.txt file listing important pages and a summary for AI crawlers.

7. Homepage lacks H1 heading Medium

   The homepage uses H2 and H3 tags but no H1, which is an unusual structural omission that may confuse crawlers about page hierarchy.

   What to change: Add a descriptive H1 tag to the homepage that includes the brand name and primary offering.

8. Product listing pages lack structured data Medium

   The Gen 4 webbing page has no Product schema, reducing AI understanding of product details.

   What to change: Add Product schema to all product listing and detail pages.

9. No historical snapshots in Wayback Machine Low

   The site has no archived snapshots, limiting AI models' ability to verify historical claims.

What's working

• All major AI crawlers receive full content — GPTBot, ClaudeBot, and other AI crawlers get a 200 response with identical content to browser baseline, with no blocking or JS shells.
• Robots.txt allows all AI bots — The robots.txt has no AI-bot-specific disallow rules, only blocking backend paths.
• Organization and Store schema with correct address — JSON-LD includes Organization and Store types with accurate address and phone number.
• Contact and Click & Collect pages resolve — The contact-us and clickandcollect pages return 200 with content, providing basic business information.
• Social media profiles linked in footer — Facebook, Instagram, and YouTube links are present, offering potential external signal sources.