The agent gathered a partial but functional understanding of isits AG by combining web search results with limited direct site access. The site's German language, heavy JavaScript rendering, and lack of publicly visible pricing information created significant friction; the agent had to piece together offerings, target audience, and differentiation from search snippets and inferred prior knowledge rather than from well-structured, machine-readable pages. The core deliverable (what they do, who it's for, how they differ) was satisfactorily assembled, but a critical gap remained: pricing could not be found anywhere.

• ›Web search snippets (steps 13, 18, 21, 22) were the primary content-bearing sources; they surfaced landing page titles, URL structure hints, and external references (MastersPortal, eurobits.de, ESCAR, CYBICS) that the agent used to infer organizational structure and positioning.
• ›Direct site fetches (steps 1, 7, 8, 10, 19) returned Next.js-rendered HTML with truncated content and no meaningful text extraction—the site's JavaScript-heavy architecture and German-only interface made it machine-hostile. The agent could not extract actual page content from any direct fetch.
• ›Pricing information was completely absent from both web search results and direct site access; the agent correctly flagged this as unfetchable rather than claiming to have found it.
• ›The agent was forced to rely on prior knowledge and web search inferences to construct answers about their Master's program (Ruhr University partnership), conference portfolio (ESCAR, CYBICS), and accreditations—these were not clearly surfaced on the site itself but were discoverable through external references.
• ›The site's navigation structure (URLs like /unternehmen, /seminare, /zertifikatsabschluesse) was navigable by URL guessing, but page content remained inaccessible due to JavaScript rendering and truncation.