The agent successfully assembled a comprehensive explanation of NinjaTrader by combining content from the homepage, pricing pages, and futures landing pages—but had to rely heavily on prior knowledge and web search to fill critical gaps. The site's information architecture is sparse; direct navigation attempts (features, advantages, platform, about, trading, products) all returned 404s, forcing the agent to search externally for competitive positioning, ownership details, and detailed fee breakdowns that should have been on-site.

• ›Steps [1], [2], [11], and [14] provided direct content: homepage tagline ('trade stocks you love, futures-style'), pricing tier structure ($0/$99/$1,499), futures broker positioning, and commission tables. These were the only on-site sources that contained actionable information.
• ›The site lacks dedicated pages for features, platform overview, about/company info, and advantages—all standard discovery pages. The agent had to guess at URLs [3–8] and got 404s, then pivoted to web search [9], [12], [13], [15–21] to fill gaps on pricing tiers, competitive differentiation, target audience, and Kraken ownership. This forced 12 of 21 steps to be search-based rather than site-native.
• ›Critical gaps not clearly surfaced on-site: (a) Kraken acquisition and ownership structure, (b) all-in fees (commissions + exchange/clearing fees), (c) $35/month inactivity penalty, (d) desktop-only/Windows-only limitations, (e) demo data delay after 14 days, (f) practical minimum funding, (g) learning curve. The agent had to synthesize these from external sources and prior knowledge (45% prior knowledge, 18% web search, 36% prior artifact).
• ›The site is moderately navigable for pricing and basic product positioning (homepage and /pricing pages load and contain structured data), but poorly navigable for comparative evaluation, company context, and hidden terms. Information density is low on core pages; most substantive answering required leaving the domain.