The agent successfully assembled a comprehensive overview of Gresham Tech's business by combining direct site exploration with supplementary web search. The homepage and product pages (reconciliation, solutions/Prime EDM) provided core positioning and product names, while the About page and search results filled gaps on company history and competitive context. However, the site itself is poorly navigable for pricing and detailed product differentiation—critical information the task required—forcing the agent to rely on prior knowledge and web search rather than on-site content.

• ›Step [1] homepage provided high-level positioning (meta description: 'Reconciliation software and tools, post trade automation and regulatory reporting solutions') but lacked product taxonomy, pricing, or competitor comparisons needed to fully answer the task.
• ›Steps [8], [12], [13], [15] revealed individual product pages (Prime EDM, Control Cloud, reconciliation) with feature descriptions, but these were scattered across different URL patterns and not linked from a clear product hub. The agent had to guess URLs rather than follow navigation.
• ›Step [21] (About/Our Journey) was not discoverable from standard site navigation (/about returned 404 in step [5]); the agent only found it via web search referral. This critical page for understanding company positioning was hidden.
• ›Pricing information was completely absent from the site (step [4] /pricing → 404). The agent had to note this as a major gap and worked around it by citing industry norms for enterprise SaaS.
• ›Competitive differentiation and alternative product mentions (SmartRecon, AutoRek) came entirely from web search (steps [27]), not from the site itself. Gresham's own pages do not position against competitors.
• ›The site showed inconsistent URL routing (e.g., /products redirects to /solutions per step [3] note) and a recent rebrand from 'Gresham Technologies' to 'Gresham' (with Alveo merger) that was only discoverable via press releases and external sources, not clear on-site.
• ›62% of the agent's information came from previous artifacts (likely cached search results and prior knowledge of the fintech space), indicating the site alone was insufficient to answer the task comprehensively.