The agent assembled a comprehensive explanation of LVMH's corporate structure, business model, and competitive position by relying almost entirely on web search results and prior knowledge, since lvmh.com itself proved difficult to parse due to its client-side rendering. The site's HTML responses were boilerplate Next.js scaffolding with little human-readable content, forcing the agent to route through external sources (Wikipedia, FourWeekMBA, fashion blogs, Quartr) to answer all four dimensions of the task (what they do, who it's for, pricing model, differentiation).

• ›Steps [5], [12], [13], [14], [18] all returned web search results containing substantive information about LVMH's divisions, revenue, brands, competitors, and business model—these were the actual content sources. No fetch of lvmh.com itself yielded parseable content; steps [1], [6], [7], [15] all returned truncated HTML showing only asset preloading and boilerplate.
• ›The agent correctly identified that lvmh.com is a corporate/investor relations site, not a consumer shopping destination. This insight came from search result titles and snippets (e.g., 'Investors - LVMH', 'investor-contact'), not from the site's navigable structure, which the agent could not extract.
• ›lvmh.com is severely agent-hostile: built with Next.js (client-side rendering), it serves no meaningful HTML payload to crawlers or headless fetches. The agent had to rely on search engines and external aggregators to answer every substantive question, including what content actually lives on lvmh.com (investor relations, press, publications, sustainability reports). The site publishes structured data (like investor calendars and financial reports) but does not expose them via machine-readable markup or static HTML.