The agent successfully assembled a comprehensive overview of The Finance Brokers by combining direct site fetches with web search results, generating a usable explanation of their business model, target audience, and positioning. However, the site itself was poorly structured for agent navigation—most dedicated product pages returned 404s, forcing the agent to rely on prior knowledge (42%), web search (16%), and fragments from homepage meta descriptions and guide pages to answer questions about pricing, fees, and eligibility criteria that should have been explicitly documented on the site.

• ›Content-bearing steps were limited: the homepage [1], about page [3], services page [7], blog [9], contact [10], and guides index [30] returned 200s, but their HTML responses were truncated in the trajectory and did not expose structured product or pricing information. The agent had to infer most details from web search snippets [16] rather than from on-site resources.
• ›Pricing information was scattered and incomplete. Step [19] fetched a 'development-finance-rates' guide (URL found in search [16]), which the agent cited as a source, but the trajectory does not show the actual content returned. The agent published specific rates (8.90% for unsecured loans, 7.25% for secured, etc.) without clear evidence of where these came from—suggesting they were sourced from prior knowledge or search snippets, not from dedicated on-site pages.
• ›Dedicated product pages (bridging-loans [26], development-finance [27], commercial-mortgages [28]) all returned 404s, indicating the site lacks traversable product documentation. The agent had to work around this by relying on the services index page [7] and external search results rather than following a clear information architecture.
• ›The site's navigation structure is opaque to agents. Multiple intuitive URLs (/pricing [5], /how-it-works [4], /faq [12], /apply [31]) returned 404s, suggesting either dynamic routing or incomplete page inventory. This forced the agent to guess at URL patterns and fall back on search.
• ›No machine-readable structured data (schema.org, JSON-LD) or explicit fee/pricing disclosures were exposed in the fetched responses. The agent had to hand-assemble answers from fragments, making it difficult to verify accuracy or confidence in the final claims about 'no upfront fees' or specific rate structures.