The agent assembled a comprehensive explanation of Fastly by relying almost entirely on external search results and prior knowledge, since the company's own website is heavily JavaScript-rendered and did not expose content in a fetchable form. The task was satisfied: the agent delivered clear answers on what Fastly does, who it's for, how it's priced, and how it differs from competitors—but this required working around an inaccessible primary source.

• ›Steps [1], [2], [3], [4], [5], [10] fetched Fastly's own domain but returned only JavaScript scaffolding with truncated HTML, making them non-content-bearing; the agent explicitly noted this barrier at step [6] and pivoted to web search.
• ›All five content-bearing steps ([7], [8], [11], [13], [14]) were web searches that returned third-party analysis sites (TechnologyCounter, SaaSworthy, Vendr, PeerSpot, SourceForge, and Fastly's docs site). None of these were discovered via navigation from fastly.com itself; the agent guessed their existence or reached them via search.
• ›The pricing details (bandwidth tiers, regional rates, minimum spend, free trial credits) came from aggregator/analyst sites, not from Fastly's pricing page, which was not accessible. The comparison table and differentiation claims were sourced from PeerSpot and SourceForge comparisons, not from Fastly's own positioning.
• ›The agent's final response cites 9 sources; all are external third-party resources or Fastly's docs subdomain—zero citations point to fastly.com's main pages, confirming that the primary domain failed to serve content to an AI agent.
• ›The agent's honesty about friction (section 'What Was Confusing or Missing') directly reflects the navigability gap: JavaScript rendering, lack of pricing transparency, missing comparison guides, and regional pricing ambiguity were all barriers imposed by the site's structure and content choices.