chapeldown.com — AI Site Grade

Analysis

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

---

Chapel Down — AI-Visibility Audit

The homepage is fully open to every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User all return 200 with identical 358KB content), yet every single sub-page — product pages, the blog, the "Our Story" page, FAQs, and even the agents.md and /.well-known/ucp endpoints — returns 403 from Cloudflare's JS challenge wall, even to a standard browser UA from this IP range. The site is a single-page AI facade: crawlers see the homepage but nothing beyond it.

Crawler Access

The robots.txt is a Shopify default with User-agent: * Allow: / and no AI-bot-specific rules. No AI crawler is explicitly blocked in the robots file. However, Cloudflare's managed challenge is the real gate: all 11 tested UAs (including Browser) get 403 on /products/brut, /pages/our-story, /blogs/journal, and /agents.md. The homepage alone passes through — likely because it is cached at the edge (cf-cache-status present, etag with page_cache). The llms.txt does return 200 with a sophisticated agent-commerce protocol document (UCP/MCP endpoints, Shopify skill instructions), but the actual endpoints it references (/.well-known/ucp, /api/ucp/mcp) are behind the same Cloudflare wall.

Content & Schema

The homepage has zero JSON-LD schema of any type — no Organization, Product, BreadcrumbList, WebSite, or FAQPage markup. This is a significant gap for a premium wine brand that sells itself as "England's leading winemaker." The homepage text is rich (1,043 words) with product listings, reviews, and a blog feed, but the blog posts themselves (dated May 2026) and all product detail pages are inaccessible to crawlers. The heading structure is flat: one H1 ("VISIT OUR HOME") and four H2 sections. There is an FAQ section on the homepage but it is not marked up with FAQPage schema.

Cold-Knowledge Gap

The LLM knows Chapel Down as "one of the largest and most recognized English wine brands," citing royal-event service (Prince Harry and Meghan Markle's wedding reception), British Airways supply, a £10M vineyard investment in 2024, and a 2023 label-redesign controversy. The site itself never mentions the royal wedding, British Airways, or the investment. The site positions itself as "England's leading winemaker" — a claim the LLM does not independently echo (it says "one of the largest"). The Kit's Coty premium range, which the site heavily promotes, is mentioned only in passing by the LLM. The "Three Graces" still wine blend the LLM cites as "well-regarded" does not appear on the homepage at all.

External Signals

The homepage links to Instagram, Facebook, Twitter, LinkedIn, and TripAdvisor — but no press mentions, awards pages, or third-party review aggregators are surfaced on the site itself. The homepage does display an International Wine Challenge 2026 Gold Medal claim for the Rosé, but there is no linked press release or award verification page. The blog references partnerships with The Jockey Club, Chelsea Flower Show, and Pub in the Park festival, but these posts are behind the Cloudflare wall.

Surprising Findings

The site is running a Shopify storefront with a sophisticated AI-agent commerce protocol (llms.txt references UCP v2026-04-08, MCP endpoints, and a Shop skill installation path) — yet the actual product catalog, blog, and brand story pages are inaccessible to the very AI agents the protocol is designed to serve. The agents.md file (referenced in robots.txt as "Agent instructions") returns 403. The /.well-known/ucp discovery endpoint returns 403. The site has built an on-ramp for AI commerce agents but left the gate locked.

Findings

1. Cloudflare JS challenge blocks all AI crawlers from sub-pages High

   Every sub-page (product pages, blog, 'Our Story', FAQs, agents.md, .well-known/ucp) returns 403 to all tested AI crawlers and even a standard browser UA. Only the homepage passes through, likely due to edge caching.

   What to change: Disable the JS challenge for AI crawler user agents (GPTBot, ClaudeBot, PerplexityBot, etc.) on sub-pages, or serve static HTML versions of key pages to bots.

2. Homepage has zero JSON-LD schema markup High

   The homepage contains no structured data of any type (Organization, Product, BreadcrumbList, WebSite, FAQPage), missing a key opportunity for AI visibility and rich search results.

   What to change: Add JSON-LD schema for Organization, WebSite, BreadcrumbList, Product (for featured wines), and FAQPage (for the FAQ section) to the homepage.

3. llms.txt references endpoints that are blocked by Cloudflare High

   The llms.txt file (200 OK) describes a sophisticated AI-agent commerce protocol with UCP/MCP endpoints, but those endpoints (/.well-known/ucp, /api/ucp/mcp) return 403, making the protocol unusable for AI agents.

   What to change: Allow AI crawler access to /.well-known/ucp and /api/ucp/mcp endpoints, and ensure agents.md is accessible.

4. Blog and product detail pages are inaccessible to crawlers High

   The blog (/blogs/journal) and product pages (e.g., /products/brut) return 403 to all bots, preventing AI systems from indexing product descriptions, reviews, and editorial content.

   What to change: Remove the JS challenge for AI crawlers on blog and product pages, or serve static HTML versions.

5. Site does not surface press mentions, awards, or third-party reviews Medium

   The homepage links to social media but lacks links to press coverage, award verification pages, or third-party review aggregators, reducing external signal strength for AI systems.

   What to change: Add a press page or awards section with links to external articles and verification pages.

6. LLM knowledge includes key brand facts not present on the site Medium

   The LLM knows about Chapel Down's royal wedding service, British Airways supply, and £10M investment, but the site does not mention these, creating a disconnect between brand reputation and on-site content.

   What to change: Add a 'Press' or 'About' section that highlights royal wedding service, British Airways partnership, and recent investments.

7. Homepage FAQ section lacks FAQPage schema Medium

   The homepage includes an FAQ section but it is not marked up with FAQPage structured data, missing an opportunity for AI-enhanced visibility.

   What to change: Add FAQPage JSON-LD schema to the FAQ section on the homepage.

8. Homepage heading structure is flat with only one H1 Low

   The homepage has a single H1 ('VISIT OUR HOME') and four H2 sections, which may limit semantic clarity for AI crawlers.

   What to change: Use a more descriptive H1 and add H2/H3 hierarchy for sections like products, reviews, and blog feed.

9. robots.txt has no AI-specific rules Low

   The robots.txt is a Shopify default with no explicit rules for AI crawlers, leaving access control entirely to Cloudflare's challenge.

   What to change: Add explicit Allow/Disallow rules for AI crawlers in robots.txt to complement Cloudflare settings.

What's working

• Homepage is fully accessible to all major AI crawlers — The homepage returns 200 with full content to GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and ChatGPT-User, providing a strong entry point for AI systems.
• llms.txt file is present and well-structured — The llms.txt file (200 OK, 4001 bytes) provides a sophisticated agent-commerce protocol with UCP/MCP endpoints and Shopify skill instructions, demonstrating forward-thinking AI readiness.
• Homepage contains rich, descriptive content — The homepage has 1,043 words with product listings, reviews, and a blog feed, providing substantial context for AI crawlers.
• Social media profiles are linked from the homepage — Links to Instagram, Facebook, Twitter, LinkedIn, and TripAdvisor provide external signal pathways for AI systems.
• Sitemap is accessible and contains 80 URLs — The sitemap.xml returns 200 and lists 80 URLs, providing a clear content inventory for crawlers.