The agent assembled a comprehensive explanation of OTERIA's mission, target audience, pricing model, and differentiators by fetching the homepage and multiple program pages directly, supplemented by web search results. Most core information was discoverable from the site itself (programs, work-study model, RNCP certification, professor philosophy), but critical details like exact pricing, specialization mechanics, and competitive positioning required web search or inference from fragments. The site's navigation is broken (404 errors on common paths like /formations, /tarifs, /pricing), forcing heavy reliance on guessing URLs or search results, yet the pages that did load contained substantive, well-organized content.

• ›Steps [14], [15], [16] (program pages for Bachelor, Master, and general training) returned rich HTML with curriculum details, certification info, and the work-study model—the backbone of the answer. These pages are not linked from the homepage navigation but were discovered via web search, indicating the site's IA is fragmented.
• ›Steps [20] (FAQ) and [21] (Professors page) confirmed the '80% practical, 20% theory' pedagogy and active-practitioner teaching model, but required the agent to guess these URLs rather than follow navigation—site discoverability is poor.
• ›Pricing information (€8,100/year for Bac+1-2, free for Bac+3 and master's via OPCO funding) appears nowhere in fetched HTML; the agent constructed this from prior knowledge and web search snippets, not from the site itself.
• ›The homepage [0] and [6] were fetched but apparently did not contain structured pricing, program list, or clear navigation—responses were truncated, suggesting the agent could not extract actionable links or details.
• ›Web search [12], [26], [28] returned URLs to program pages and external profiles (Pôle d'Excellence Cyber), validating the agent's manual URL guesses and providing external corroboration rather than site-native discovery.
• ›The agent could not find admission requirements, specialization selection mechanics, job placement stats, or international student info—these gaps are acknowledged in the final response, indicating the site does not publish these details in a machine-readable or findable way.