The agent assembled a comprehensive explanation of France Student by combining the official homepage (which was inaccessible as a rendered React app) with a detailed third-party article from Slash Magazine. The site's JavaScript-heavy architecture forced the agent to rely on web search and external sources rather than direct navigation, making discovery work around a poorly agent-readable interface. Despite these friction points, the agent successfully answered all four core questions (what they do, who it's for, pricing, and differentiation) by synthesizing fragments from cached homepage metadata and external reporting.

• ›Step [1] (homepage) returned only HTML boilerplate and truncated meta tags—the actual interface content was in JavaScript bundles that didn't execute. The meta description promised 'free digital tools' but no inventory or feature list was visible in the fetch response.
• ›Step [11] (Slash Magazine article) was the only richly informative fetch—it provided mission context ('fight student precarity'), explicit tool names (Canva, Adobe), eligibility (all French students), and the non-profit positioning that became central to the agent's differentiation narrative.
• ›The agent relied heavily on web search (80% of sourcing per metadata) to route around the inaccessible site. Searches for specific tools, pricing, and comparisons returned snippets and external coverage rather than first-party content, indicating France Student does not publish structured, discoverable information on its own domain.
• ›Step [6] (deals page) and attempted Step [4] (learn subdomain) both failed or returned the same truncated homepage, showing the site does not have distinct, agent-accessible landing pages for key sections.
• ›The agent correctly noted confusion about the commerce layer ('my.francestudent.org/store' and renewal mechanics) because these details were not exposed by any successful fetch—they existed only as URLs found in search results but could not be visited.
• ›The agent's final explanation relies on explicit citation of only two sources that were fetched: the official homepage (for branding) and the Slash Magazine article (for substantive details). This asymmetry reveals the site itself is poorly positioned as a primary source.