The agent assembled a comprehensive explanation of McMillan Education International by combining one successfully fetched homepage, four internal pages that returned valid HTML (universities, boarding-schools, team, contact), and six web searches. The site itself was poorly navigable—most direct requests timed out, forcing heavy reliance on search results and prior knowledge to reconstruct the company's purpose, positioning, and services. The agent successfully answered what they do and who it's for, but flagged pricing as opaque (not publicly disclosed on the site).

• ›Only the homepage [3] and four internal pages [13, 14, 16, 17] actually returned content; all other direct fetches to key pages (about, pricing) timed out, indicating serious site performance or availability issues.
• ›The agent retrieved actual HTML from /universities, /boarding-schools, /about/team, and /contact but the responses were truncated and did not contain readable content—the agent relied on URL structure and page titles inferred from search results rather than page content.
• ›Six web searches [9, 10, 19, 22, 23, 25] provided the bulk of the answer: they surfaced the About page URL, team page URL, and contextual information about services, experience (70 years, 20+ year average consultant tenure), and competitor landscape. Search snippets, not the site's own HTML, carried the factual payload.
• ›Pricing information was entirely absent from the website and had to be inferred from the sister company (McMillan Education US) pricing ($15,000+) and the agent's instruction to 'contact for custom quote'—this was explicitly called out as confusing/missing.
• ›The site does not publish pricing tiers, detailed service breakdowns by package, or clear differentiation from the US-focused sister company in a fetchable form; the agent assembled the explanation from search-surfaced metadata, prior knowledge, and deduction rather than from the site's own content architecture.