The agent successfully gathered enough information from three core pages to describe Frontier Labs' business model, target audience, and research themes. However, the site itself is sparse on evidence—no client names, case studies with outcomes, team details, pricing, or competitive positioning—forcing the agent to report what the site *doesn't* say as much as what it does.

• ›The site explicitly labels its only featured system (Intelligence Fusion Platform) as a design concept and not a real product or case study, yet provides a link ('View system architecture') that could mislead users into thinking a working system exists.
• ›Critical business information is entirely absent from the fetched pages: no team/leadership, no client references, no measurable outcomes, no delivery process, no engagement models, and no pricing—making it impossible for prospects to assess credibility or fit without direct contact.
• ›The distinction between 'Work' and 'Labs' is unclear from the site itself; the agent had to infer that Labs is R&D and Work is portfolio, but Work showed only the same illustrative concept, not actual delivery examples.
• ›The site relies solely on contact forms and email for inquiry, with no intermediate content (e.g., case studies, team bios, service tiers, technical architecture docs) to help prospects self-qualify or understand engagement depth before outreach.
• ›The agent did not fetch /capabilities, /company, or /insights pages, which could have filled some gaps, but the final response appropriately flagged this limitation; the discovery was driven primarily by homepage and two second-level pages.