The agent successfully gathered substantive information about Do Good Jobs' mission, pricing, and positioning by navigating the site's core pages. The site was reasonably navigable but presented some friction: the homepage advertises 'not yet live' status while still hosting pricing and About pages, creating domain confusion that the agent correctly identified but could not fully resolve from on-site sources.

• ›Step [0] and [1] revealed that the homepage itself carries the core value proposition and sector focus, making the landing page content-rich, but step [1] also showed that actual pricing details are gated behind a /pricing link rather than displayed on the homepage—requiring an extra fetch.
• ›Step [2] (the /pricing page) provided complete pricing detail in structured, fetched form—job listing tiers, add-ons, and sponsorship packages with explicit NZD amounts—making this the most content-dense step. The agent used this as the sole basis for the Pricing section.
• ›Step [5] (the /about page) supplied mission, values, founding context, and team structure—all content the final response drew from to explain 'how they're different' and who founded the org. This page was critical for the positioning narrative.
• ›The agent correctly flagged but could not resolve the dogoodjobs.org vs. DoGoodJobs.co.nz discrepancy from on-site content alone; the homepage banner mentioning the site is 'not yet live' was exposed by the fetch, but no page explained the relationship between the two domains—a gap the agent attributed to model inference rather than retrieved fact.
• ›Steps [3] and [4] were navigation/structure discovery (impact categories, nav menu, job count) rather than content-bearing for the final answer; they helped the agent map the site but did not contribute facts to the summary.