The agent successfully compiled a comprehensive explanation of Swan Bitcoin by fetching the homepage, pricing/fees page, about page, and Swan Private product page, supplemented by a web search for competitor context. The site was reasonably navigable for discovery purposes, with key information published in structured JSON endpoints and standard pages, though the agent had to rely on prior knowledge and web search to fully contextualize Swan's positioning relative to alternatives.

• ›Steps [4] and [5] yielded structured JSON data from DatoCMS endpoints (/page-data/fees/page-data.json and /page-data/about/page-data.json) containing pricing, custody options, and company mission — the site publishes this in machine-readable form, making extraction straightforward for pricing and feature details.
• ›Step [6] (homepage JSON) confirmed the product positioning ('Build Bitcoin Wealth with Personalized Service from a Dedicated Account Representative') and promo status, but the homepage HTML responses were truncated in the trajectory, limiting what the agent could have extracted from rendered marketing copy.
• ›Step [12] and [13] revealed a secondary product tier (Swan Private/Swan Wealth) for high-net-worth clients, expanding the agent's understanding of the target market — this was discoverable via navigation but not prominent on the main pages fetched.
• ›Step [14] (web search) was necessary because the site itself does not contain a competitive comparison or alternatives section; the agent had to leave the domain to contextualize Swan against Coinbase, River, Unchained, and Strike.
• ›The agent encountered multiple 404s ([3], [7], [8], [9], [10], [11]) when guessing at logical page paths (/how-it-works, /features, /compare, /for-individuals), indicating the site's navigation surface is limited or not discoverable via standard URL patterns — the agent had to rely on prior knowledge to guess these paths rather than following links from the homepage.