The agent assembled a largely complete picture of NTB through external research (Wikipedia, case studies, industry news) rather than from ntb.no itself. The site's two main domains—ntb.no (news-focused, login-gated) and ntb.tech (technology marketing)—did not surfacing enough operational detail to answer the task; the agent relied on prior knowledge and web search to fill critical gaps about business model, ownership, pricing, and competitive positioning. Pricing information could not be found anywhere publicly.

• ›Steps [12], [18], [19] (Wikipedia and industry sources) returned the core facts about NTB's history, ownership structure (media-company-owned cooperative), and geographic focus—information not clearly available on ntb.no or ntb.tech themselves.
• ›Step [13] (Superdesk case study) and Step [14] (news acquisition article) revealed NTB's technology acquisitions and platform names, which were not prominently detailed on the official site's marketing pages.
• ›The official site was poorly structured for discovery: ntb.no appears as a news aggregator/portal (requires login to see business info), while ntb.tech is a separate B2B marketing site focused on technology buyers. Neither clearly explains what NTB *is* to a first-time visitor or why it differs from competitors.
• ›Pricing was completely absent from all public pages ([8] fetched ntb.tech/en/pricing but returned no accessible price data). The agent had to note this as a gap rather than retrieving it.
• ›The agent's fetches of ntb.no/teknologi [2], ntb.no/kontakt [4], and ntb.tech/en/solutions [16] returned HTML but yielded minimal extractable text in the trajectory—suggesting the pages are JavaScript-heavy or behind paywalls, not agent-friendly.
• ›The site's structure forces visitors to choose between two separate properties (news site vs. tech platform) with no clear integration or explanation of how they relate, creating friction for understanding the unified business.