kalypso.com — AI Site Grade

Analysis

I have enough data to write a thorough audit. Let me compile.

Kalypso blocks GPTBot, OAI-SearchBot, PerplexityBot, ChatGPT-User, and anthropic-ai at the Cloudflare edge with 403s, while ClaudeBot and Google-Extended get full 200 access — a selective AI blockade that directly contradicts the robots.txt file, which explicitly disallows ClaudeBot but allows the OpenAI family.

Crawler Access

The robots.txt at kalypso.com/robots.txt is unusually aggressive: it disallows 41 user-agents including ClaudeBot, Bytespider, CCBot, cohere-ai, and meta-externalagent. However, GPTBot, OAI-SearchBot, Google-Extended, and PerplexityBot are not listed in the file at all — the wildcard User-agent: * with Allow: / governs them. The robots.txt also carries a non-standard Content-signal: search=yes,ai-train=no directive, an experimental signal that most crawlers ignore.

In practice, compare_bot_access on the homepage reveals a different blockade at the Cloudflare edge: GPTBot, OAI-SearchBot, ChatGPT-User, PerplexityBot, Perplexity-User, Bytespider, and anthropic-ai all receive HTTP 403 with a 4.5 KB Cloudflare block page. ClaudeBot and Google-Extended return 200 with the full 273 KB page — meaning the Cloudflare WAF rules override the robots.txt intent for ClaudeBot (which is disallowed in the file but allowed at the edge) and enforce a block on OpenAI/Perplexity that the robots.txt never requested. This is a mismatched, inconsistent posture: the robots.txt says one thing, the WAF does another.

Cold-Knowledge Gap

The LLM prior knows Kalypso as a consulting firm focused on product lifecycle management (PLM), IoT strategy, and innovation management, often leveraging PTC Windchill. It recalls the 2018 Rockwell Automation acquisition and describes Kalypso as having "reduced standalone brand visibility" post-acquisition. The actual site tells a different story: Kalypso now positions itself as Rockwell Automation's digital service arm driving the shift "from industrial automation to autonomy," with heavy emphasis on generative AI copilots for industrial operations, AI-powered factory of the future, and data science services. The site mentions PTC only as one partner among many (Microsoft, NVIDIA, Oracle, Collibra). The cold knowledge is stale by roughly 5-7 years — it describes pre-acquisition Kalypso, missing the current GenAI and autonomous-manufacturing narrative entirely.

Schema Posture

Every page carries a rich Corporation schema block with full address (1201 S 2nd St, Milwaukee, WI 53204), social profiles, logo, and a detailed description. However, the schema type is always Corporation — never ProfessionalService, ConsultingBusiness, or Organization. Service pages use @type: Service but the homepage and about page use Corporation, which is semantically imprecise for a consulting firm. No FAQPage, Article, HowTo, or Product schema exists anywhere. The breadcrumb BreadcrumbList schema is present on all subpages. No llms.txt exists (404).

External Signals

The press releases page shows steady activity through September 2025, with mentions of Gartner Life Sciences Guide recognition, Michelin partnership, and PTC Digital Manufacturing Partner of the Year awards. The site references 4,388 projects delivered, 1,000+ brands served, 100+ countries. DNS is hosted on Azure DNS with Cloudflare CDN (A records point to Cloudflare IPs). The site runs on Craft CMS (x-powered-by: Craft CMS). No significant Reddit or third-party review presence was found in search results — the brand's external AI footprint is thin beyond its own press releases and LinkedIn activity.

Findings

1. Cloudflare edge blocks contradict robots.txt directives for AI crawlers High

   The robots.txt disallows ClaudeBot but allows GPTBot, OAI-SearchBot, and PerplexityBot via wildcard. However, Cloudflare WAF returns 403 for OpenAI and Perplexity bots while allowing ClaudeBot (200). This mismatch creates an inconsistent AI visibility posture.

   What to change: Align Cloudflare WAF rules with robots.txt intent: either allow GPTBot/OAI-SearchBot/PerplexityBot at the edge or update robots.txt to disallow them. Remove the non-standard Content-signal directive.

2. LLM cold knowledge is 5-7 years out of date High

   The LLM prior describes Kalypso as a PLM-focused consulting firm leveraging PTC Windchill, pre-acquisition by Rockwell Automation. The actual site positions Kalypso as Rockwell's digital service arm with heavy emphasis on generative AI copilots and autonomous manufacturing. This gap means AI assistants will provide outdated information about the firm.

   What to change: Publish an llms.txt file with current positioning, key services, and recent press releases. Ensure high-authority pages (homepage, about, services) are crawlable by AI bots and contain up-to-date descriptions.

3. No llms.txt file published Medium

   The site returns 404 for /llms.txt, missing an opportunity to provide structured guidance to AI crawlers about which pages to index and how to interpret the site's content.

   What to change: Create an llms.txt file that lists key pages (homepage, services, about, press releases) and provides a brief summary of the firm's current focus.

4. Homepage and about page use Corporation schema instead of ProfessionalService Medium

   The schema type on the homepage and about page is always 'Corporation', which is semantically imprecise for a consulting firm. 'ProfessionalService' or 'ConsultingBusiness' would better represent the organization type to knowledge graphs.

   What to change: Change the @type on the homepage and about page from 'Corporation' to 'ProfessionalService' or 'ConsultingBusiness'.

5. No FAQPage, Article, or HowTo schema on relevant pages Medium

   Despite having service pages and a press releases section, the site does not use FAQPage, Article, or HowTo schema. This limits the ability of AI crawlers to extract structured Q&A or article content for featured snippets.

   What to change: Add FAQPage schema to service pages where common questions are addressed, and Article schema to press releases.

6. Minimal external presence on Reddit and third-party review sites Medium

   Web searches for Kalypso on Reddit and general review sites returned zero results. The brand's external AI footprint is thin beyond its own press releases and LinkedIn, reducing the volume of third-party signals that AI models can reference.

   What to change: Encourage client testimonials on third-party platforms like G2 or Clutch, and engage in relevant industry forums to build external citations.

7. Viewpoints page returns 404 Low

   The URL /about/viewpoints returns a 404 page, indicating a broken link or removed content. This may have been a resource for thought leadership content that is no longer accessible.

   What to change: Restore the viewpoints page or set up a redirect to a relevant section (e.g., /about/press/press-releases).

What's working

• Consistent Corporation schema with full address and social profiles — Every page carries a rich Corporation schema block with complete address, social profiles, logo, and description, providing strong structured data for knowledge graphs.
• BreadcrumbList schema on all subpages — All subpages include BreadcrumbList schema, aiding navigation understanding for crawlers.
• Service schema on service pages — Service pages use @type: Service schema, providing specific structured data for the services offered.
• Active press releases through September 2025 — The press releases page shows steady activity with recent mentions of Gartner recognition, Michelin partnership, and PTC awards, signaling ongoing relevance.
• Cloudflare CDN with fast response times — The site uses Cloudflare CDN, providing fast global delivery and DDoS protection.
• Sitemap with 63 URLs and index enabled — The sitemap contains 63 URLs and is properly indexed, aiding crawler discovery.