The agent successfully assembled a comprehensive overview of work21.nl's business model, positioning, and services, but encountered significant friction due to the site's heavy JavaScript rendering and lack of transparent pricing information. The agent relied heavily on web search results and prior knowledge to fill gaps left by an inaccessible website—direct page fetches returned compressed HTML with minimal readable content, while the WordPress REST API provided only partial structural data. Key findings came from search results and external business profiles rather than from the site's own pages.

• ›Web search steps [19], [27], [28], [30], [34] were the primary content-bearing sources. These search results surfaced work21's service categories (change management, strategic transformation, masterclasses), their methodology ('Verander-DNA'), and external company profile data that the agent then synthesized into its answer. Direct page fetches (steps [16], [17], [21], [22]) returned only truncated, heavily minified HTML unsuitable for parsing.
• ›The site did not surface pricing information through any accessible route—no dedicated pricing page, no REST API endpoint, no search snippet contained cost data. The agent explicitly flagged this as 'unclear/confusing' and correctly concluded the pricing model is custom/quote-based. This represents a significant discovery barrier: a prospect cannot self-educate on cost.
• ›The WordPress REST API (steps [24], [25], [31], [33]) provided page titles and partial content but responses were truncated mid-fragment, making it impossible to extract methodology details or service package boundaries from the API alone. The agent had to resort to web search and prior knowledge to answer 'how it's different' and 'how it's priced'—information that should be directly available on the site.
• ›The site's core differentiator ('Verander-DNA') is mentioned in search results and external HR community profiles but is poorly explained on the actual website itself. This is a positioning failure: the agent's final summary of this concept relies on inference and external sources, not on clear documentation from work21's own pages.
• ›All direct page URL guesses for standard pages failed (steps [5], [6] for /pricing and /about returned 404s), forcing the agent to discover structure via REST API and search. The site uses non-standard URL patterns and relies on JavaScript navigation, making it difficult for agents to guess or crawl the navigation hierarchy.