cundall.com — AI Site Grade

Analysis

Cundall.com — AI-Visibility Audit

The site's most consequential AI-visibility problem is not a block — it is a complete absence of structured data across every page examined, combined with a silent erasure of the brand's most distinctive credential (B Corp certification) from its own website, creating a wide gap between what AI models know about Cundall and what the site actually tells them.

Crawler Access

/robots.txt and /llms.txt both return 404. The site has no robot exclusion rules and no AI-friendly content map whatsoever. compare_bot_access on the homepage reveals a split-access pattern: GPTBot and ClaudeBot receive 403 blocks (likely from the Cloudflare/WAF layer — the site runs on AWS with Route53 DNS), while ChatGPT-User, OAI-SearchBot, Google-Extended, PerplexityBot, Perplexity-User, and anthropic-ai all get 200 with full content (125KB, identical to browser baseline). Bytespider is also blocked. This means OpenAI's search and training crawlers see the site differently — ChatGPT-User can ingest content, but GPTBot (training) cannot. The sitemap.xml exists (483KB, ~500+ URLs) and is well-maintained with recent lastmod dates (May 2026), so Google can discover the full content graph.

Content & Schema Posture

Zero JSON-LD schema was found on any page: not on the homepage, /about, /services, /projects, /sectors, /about/sustainability, or /about/zero-carbon-design-2030. No Organization, ProfessionalService, WebSite, BreadcrumbList, or FAQPage schema exists anywhere. Every page uses the same generic meta description ("Cundall is a global, independent, multi-disciplinary consultancy..."). The /services and /sectors pages are thin-content shells — 42 and 49 words of visible text respectively, just a list of service/sector names with no descriptions, no schema, no structured data. The /projects page uses infinite-scroll ("Loading more") which AI crawlers cannot paginate through. The homepage has 576 words of text but no H1 beyond "Home" — the actual brand positioning is buried in the meta description and image alt text.

Cold-Knowledge Gap

The LLM queried cold knows Cundall as a founding B Corp-certified engineering consultancy (since 2015), a signatory to the One Planet Living framework, the firm behind The Edge in Amsterdam ("world's greenest office"), and a Best for the World awardee. None of these facts appear on cundall.com. A site-wide search for "B Corp" or "B Corporation" returns zero results. The Edge project page returns a 404. The "One Planet" framework is mentioned only in passing on the sustainability page as "One Planet Living Principles" — no dedicated page, no badge, no certification detail. The site mentions "Zero Carbon Design 2030" prominently but omits the B Corp status that would give instant credibility to that claim. The model's prior is richer and more specific than anything the site surfaces.

External Signals

The site runs on Craft CMS (evident from CRAFT_CSRF_TOKEN in forms) hosted on AWS (35.177.23.173, Route53 DNS). Mimecast handles email security. The robots meta tag is set to noodp (a legacy directive preventing DMOZ descriptions — largely irrelevant today). No noindex or nofollow directives were found on any page. The site has a CDN subdomain (cdn.cundall.com) serving assets with aggressive cache-busting via versioned URLs. No major external review or press results surfaced in search, suggesting limited third-party citation volume that AI engines could draw on.

Surprising Signals

The 404 on /projects/the-edge is notable — The Edge in Amsterdam is arguably Cundall's most famous project globally, and the cold model knows it well. Its absence from the live site means AI engines retrieving the site in real-time cannot connect Cundall to this signature work. The 50th anniversary microsite (/about/cundall-at-50) is content-rich but also schema-free. The site has a podcast ("Things That Really Matter") and active blog content with recent dates (May 2026), indicating good content velocity — but none of it is surfaced to AI crawlers via structured data or an llms.txt.

Findings

1. Zero JSON-LD structured data on any page High

   No Organization, ProfessionalService, WebSite, BreadcrumbList, or FAQPage schema exists on any page examined, including the homepage, about, services, projects, and sustainability pages.

   What to change: Add JSON-LD structured data for Organization, ProfessionalService, WebSite, BreadcrumbList, and relevant page types across all pages.

2. B Corp certification absent from the website High

   Cundall is a founding B Corp-certified engineering consultancy (since 2015), but no mention of 'B Corp' or 'B Corporation' exists on the site. A site-wide search returns zero results.

   What to change: Add a dedicated page or section highlighting B Corp certification, including the certification badge and details, and link to the B Corp directory.

3. GPTBot and ClaudeBot receive 403 blocks High

   GPTBot and ClaudeBot are blocked by the Cloudflare/WAF layer, while ChatGPT-User and other crawlers get full content. This limits AI training data ingestion.

   What to change: Allow GPTBot and ClaudeBot access by adjusting WAF rules or adding explicit allow directives in robots.txt.

4. The Edge project page returns 404 High

   The Edge in Amsterdam, Cundall's most famous project, returns a 404 error. AI models know about it from prior knowledge, but the site cannot confirm it.

   What to change: Restore or create a project page for The Edge with detailed content and structured data.

5. Services and sectors pages are thin-content shells Medium

   The /services and /sectors pages contain only 49 and 42 words respectively, listing names without descriptions or structured data.

   What to change: Expand each service and sector page with descriptive content, case studies, and relevant schema markup.

6. Projects page uses infinite scroll not crawlable by AI Medium

   The /projects page loads more content via infinite scroll ('Loading more'), which AI crawlers cannot paginate through, limiting discoverability of individual project pages.

   What to change: Implement traditional pagination with crawlable links or a sitemap that includes all project URLs.

7. No llms.txt file for AI-friendly content map Medium

   The site does not provide an llms.txt file, which would help AI crawlers discover key content efficiently.

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

8. Robots.txt returns 404 with no rules Medium

   The robots.txt file is missing, returning a 404. No user-agent rules or sitemap references are provided.

   What to change: Create a robots.txt file that allows all crawlers and references the sitemap.

9. Generic meta description used across all pages Low

   Every page uses the same generic meta description, missing opportunities to differentiate content in search snippets.

   What to change: Write unique, descriptive meta descriptions for each page that include key differentiators.

10. One Planet Living framework mentioned only in passing Low

    The sustainability page mentions 'One Planet Living Principles' but lacks a dedicated page or detailed explanation of Cundall's commitment to the framework.

    What to change: Create a dedicated page explaining the One Planet Living framework and Cundall's role as a signatory.

11. Homepage lacks a descriptive H1 tag Low

    The homepage H1 is simply 'Home', missing an opportunity to communicate brand positioning to search engines and AI.

    What to change: Change the H1 to a descriptive tagline such as 'Global Engineering Consultancy for Sustainable Design'.

12. Limited third-party citations and press results Low

    Web searches for Cundall reviews and awards returned zero results, indicating low external citation volume that AI engines can draw on.

What's working

• Sitemap.xml exists with 500+ URLs and recent lastmod dates — The sitemap is well-maintained with recent lastmod dates (May 2026), enabling Google to discover the full content graph.
• ChatGPT-User, OAI-SearchBot, and other key crawlers get full content — OpenAI's search and user crawlers, Google-Extended, PerplexityBot, and anthropic-ai all receive 200 with full content, enabling real-time retrieval.
• Active blog and podcast with recent content — The site has a podcast ('Things That Really Matter') and blog with recent dates (May 2026), indicating good content velocity.
• No noindex or nofollow directives on any page — All pages are indexable and followable, allowing search engines and AI crawlers to discover content freely.
• Dedicated Zero Carbon Design 2030 page with substantial content — The page contains 658 words of detailed content about Cundall's sustainability commitment, providing rich material for AI models.
• Rich history and 50th anniversary pages — The about/history and about/cundall-at-50 pages contain substantial text (897 and 610 words) detailing the firm's background and milestones.
• CDN subdomain with aggressive cache-busting for assets — Static assets are served via cdn.cundall.com with versioned URLs, improving load performance and crawl efficiency.
• Compliance page with detailed policies — The about/compliance page contains 381 words covering modern slavery, anti-bribery, and whistleblowing policies, demonstrating governance transparency.