The agent successfully assembled a comprehensive overview of Bruce Studios by combining content from the site's core pages (homepage, about, how-it-works, memberships, FAQ) with web search results to fill pricing gaps. The site is moderately navigable for discovery but pricing information is not directly exposed in fetched HTML—the agent had to rely on prior knowledge and search to construct specific price points, and many standard pricing/features pages (e.g., /pricing, /features, /plans) returned 404s.

• ›Steps [1], [5], [7], [10], [18], [19], [24] returned valid pages establishing what Bruce Studios does (multi-studio Nordic fitness membership, 1000+ partner locations, BASE/BLACK/EPIC tiers, mobile app), but pricing specifics (SEK/DKK/NOK amounts, promotion percentages) were not visible in the HTML responses—the agent inferred these from web search [22] and prior knowledge rather than from the site itself.
• ›The site's information architecture is fragmented: standard pricing routes (/pricing, /features, /plans, /membership) returned 404s, forcing the agent to discover pricing via /memberships and /en/membership pages. The site appears to use JavaScript-heavy rendering that may not expose pricing in static HTML, making it opaque to fetch-based extraction.
• ›The agent had to compensate for missing competitive positioning and feature clarity on the site by using web search [16] and [26] to contextualize Bruce against ClassPass and Swiftr. The site itself does not explicitly compare itself to alternatives or provide granular feature breakdowns (e.g., what qualifies as 'most studios' vs. 'more studios').
• ›Steps [20] (help.bruce.app FAQ) and [24] (FAQ page) were cited as sources but the trajectory shows they were fetched after the agent had already constructed most of the answer from earlier content-bearing steps, suggesting the FAQ was accessed for confirmation rather than discovery.