e-worx.co.uk — AI Site Grade

Analysis

e-worx publishes an llms.txt with a live MCP endpoint — a rare and advanced AI-access posture — yet the site ships two case studies filled with placeholder text ("This is placeholder text. To change this content...") that are indexed and sitemap-registered, directly undermining the credibility AI agents would surface.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receive a 200 status with real HTML content from the Wix-hosted Pepyaka server. No UA-based blocking exists. The robots.txt uses a catch-all Allow: / with no AI-bot-specific disallow rules, meaning every AI crawler has full run of the site. The llms.txt is present and unusually sophisticated: it documents a Wix MCP endpoint (/_api/mcp) that lets AI agents query live business details, search site content, and even trigger purchases via API calls — no scraping required.

Cold-Knowledge Gap

A frontier LLM queried cold about e-worx returned no specific knowledge — could not name services, clients, case studies, or any reputational signal. The model guessed generic "SEO, PPC, social media" services. The site itself positions as a data-driven digital sales agency with a strong Amazon marketplace and TikTok Shop specialism, plus NPD/market insights — a differentiated pitch that is entirely absent from the model's prior. The gap between the site's actual positioning (Amazon-first, TikTok Shop, full-service digital sales) and what AI models know cold (nothing) is near-total.

Schema and Content Posture

The homepage carries LocalBusiness and WebSite JSON-LD with address, phone, and URL — correct but minimal. No Service schema, no FAQPage, no Product schema, no Review schema on any page. The about page and services page have zero JSON-LD. Blog posts use BlogPosting schema with author and dates. The case study pages use only ImageObject schema — no CaseStudy or Article type. Two of the three case studies (quintessential-offices and quintessential-offices2) contain Wix placeholder text ("This is placeholder text. To change this content, double-click on the element and click Change Content.") and are indexed and listed in the sitemap. The third case study (born-in-scotland) is legitimate with real results data.

External Signals

The site has near-zero external footprint. Searches for "e-worx" combined with Glasgow, digital marketing, or reviews return no results on DuckDuckGo. The only external link found on the entire site is to the company's LinkedIn page. No Reddit threads, no press mentions, no review sites, no directory listings surfaced. The DNS shows Google Workspace mail (aspmx.l.google.com) and Wix nameservers. The site's only backlink profile appears to be self-referential. For an agency claiming to drive digital sales, the absence of off-domain signals means AI engines have no corroborating third-party data to cite.

Surprising Details

The llms.txt is a standout — it explicitly invites AI agents to connect via MCP and perform actions like booking and purchasing. This is an advanced posture that most sites in this category lack. However, the presence of two published, indexed case studies with placeholder copy undermines the quality signal an AI agent would retrieve. The blog has 10+ posts with solid word counts (2,000–3,000 words) and FAQ sections, but they are generic "What is X" explainers with no proprietary data or client-specific insights. The homepage lists "Smart Protein" as a project but the case study for it is missing from the sitemap.

Findings

1. Two indexed case studies contain Wix placeholder text High

   The case studies at /case-studies/quintessential-offices and /case-studies/quintessential-offices2 display placeholder copy such as 'This is placeholder text. To change this content, double-click on the element and click Change Content.' These pages are indexed and listed in the sitemap, damaging credibility for AI agents.

   What to change: Replace placeholder content with real case study data or remove these pages from the sitemap and index.

2. Frontier LLMs have no specific knowledge of e-worx High

   A cold query about e-worx returned no specific knowledge; the model could not name services, clients, or case studies. The site's differentiated positioning (Amazon marketplace, TikTok Shop) is absent from AI priors.

   What to change: Increase off-site signals (backlinks, reviews, press) and ensure schema-rich content is crawlable to build AI knowledge.

3. About and services pages lack JSON-LD schema High

   The about page and services page have zero JSON-LD markup. Case study pages use only ImageObject schema, missing CaseStudy or Article types. No Service, FAQPage, or Review schema exists anywhere.

   What to change: Add Service, FAQPage, and CaseStudy schema to relevant pages. Include Review schema if testimonials exist.

4. Site has near-zero external backlinks and mentions High

   Searches for 'e-worx' combined with Glasgow, digital marketing, or reviews return no results on DuckDuckGo. The only external link found is to LinkedIn. No press, reviews, or directory listings surfaced.

   What to change: Build a backlink strategy, claim directory listings, and encourage client reviews on third-party platforms.

5. Blog posts are generic explainers with no proprietary data Medium

   The blog contains 10+ posts with solid word counts (2,000–3,000 words) but they are generic 'What is X' explainers lacking proprietary data or client-specific insights, reducing differentiation for AI retrieval.

   What to change: Incorporate original research, client results, or proprietary data into blog posts to increase uniqueness and authority.

6. Smart Protein project mentioned on homepage but no case study in sitemap Medium

   The homepage lists 'Smart Protein' as a project, but no corresponding case study is found in the sitemap or on the site, creating a broken narrative for AI agents.

   What to change: Create and publish a case study for Smart Protein, or remove the mention from the homepage.

7. Homepage schema is minimal LocalBusiness only Medium

   The homepage carries LocalBusiness and WebSite JSON-LD with basic contact info, but no Service, FAQPage, or Review schema, limiting rich results for AI crawlers.

   What to change: Add Service schema listing core offerings and FAQPage schema for common questions.

8. Robots.txt has no AI-bot-specific rules Low

   The robots.txt uses a catch-all Allow: / with no disallow rules for AI crawlers. While this grants full access, it also means no rate-limiting or selective blocking is in place, which could be refined.

   What to change: Consider adding crawl-delay directives for aggressive bots or disallowing low-value paths like placeholder pages.

What's working

• llms.txt published with live Wix MCP endpoint — The site publishes an llms.txt that documents a Wix MCP endpoint, allowing AI agents to query live business details, search content, and trigger actions without scraping.
• All major AI crawlers receive 200 with real HTML — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others get full HTML content from the Wix server with no UA-based blocking.
• Robots.txt allows all crawlers with no disallow rules — The robots.txt uses a catch-all Allow: /, ensuring no AI bot is inadvertently blocked.
• Blog posts use BlogPosting schema with author and dates — Blog articles include BlogPosting JSON-LD, providing structured metadata for AI crawlers.
• One case study contains real results data — The 'born-in-scotland' case study provides legitimate results and is not placeholder content.
• Sitemap is published and contains 30 URLs — A valid sitemap is available, helping crawlers discover site content.