The agent successfully assembled a comprehensive overview of KMD by combining light direct site fetches with heavy reliance on web search results and prior knowledge. The kmd.dk website itself provided only basic metadata and navigation structure; the agent extracted most substantive details (products, market position, competitors, cloud migration strategy) from external sources like industry news, Owler, and CB Insights. The site was poorly navigable for machine extraction—it rendered via heavy JavaScript, lacked transparent pricing, and did not publish a structured product catalog or differentiation narrative.

• ›Direct fetches of kmd.dk [1, 3, 5, 6] returned only HTML headers and meta descriptions; the actual page body content was truncated and did not load, indicating JavaScript-dependent rendering that broke machine readability.
• ›A direct attempt to find pricing (/priser) returned a 404 [7], confirming that KMD does not publish pricing publicly on their site—the agent had to infer pricing models from sparse regulatory language mentioned elsewhere.
• ›The agent discovered critical details (KMD Opus, Payroll Cloud, SAP/Azure cloud migration, competitor landscape) from web search results [9–12, 14, 16] and the kmd.net English mirror [13, 15], not from kmd.dk itself. Prior knowledge (29%) and previous artifacts (71%) together supplied 100% of the synthesis; zero information came from live web search (0%).
• ›The agent correctly identified five confusing gaps: pricing opacity, JavaScript-heavy rendering, unclear product breadth, vague strategic claims, and unstated international strategy—all of which reflect genuine site design choices that impeded agent comprehension.
• ›The final answer cited external sources (Owler, CB Insights, SAP news, Intelligent CIO) rather than kmd.dk or kmd.net, signaling that the site itself does not publish the comparative, competitive, or strategic intelligence a user would need to 'understand' KMD in depth.