The agent successfully synthesized an explanation of cdrcompare.au by assembling information from the homepage, about page, compare page, and multiple web searches, since the site's JavaScript-heavy rendering and missing help/pricing pages prevented direct access to core information. The agent had to rely heavily on prior knowledge and web search results to construct a coherent answer about the service's purpose, differentiators, and business model, revealing the site is not agent-friendly for discovery-level questions.

• ›Steps [1], [3], and [7] returned Next.js-rendered HTML that was too abstract to extract actionable content; the agent could not parse the UI or feature set from raw response bodies.
• ›The agent's core insights came from web search results [9], [14], [16], and [21], which surfaced competitor context (Canstar, RateCity, Finder) and external descriptions of Open Product Compare's CDR-based model. The site itself did not publish this comparative positioning.
• ›Critical pages (/pricing, /faq, /how-it-works) returned 404 errors [4], [5], [8], leaving the agent unable to find pricing, help documentation, or operational details directly from the site.
• ›The agent inferred the business model (free, no affiliate links, API-driven) and technical differentiator (Consumer Data Right / CDR infrastructure) from Fiskil branding clues and external sources, not from explicit site content.
• ›The site's /about page [3] and /compare page [7] existed but were not machine-readable in the response; the agent had to reconstruct their meaning from metadata, search results, and domain expertise.