The agent successfully assembled a comprehensive overview of Extracadabra by combining site exploration, FAQ content, and external research. The site's information architecture made discovery difficult—critical details were scattered across fragmented FAQ pages, domain-specific subdomains, and blog content rather than consolidated landing pages. Pricing remains opaque; the agent correctly identified this as a gap and compensated by pulling financial data from third-party sources, but could not retrieve the actual employer commission rates from Extracadabra's own properties.

• ›Steps [3] (homepage) and [27] (contract types page) provided surface-level positioning and featured job categories, but lacked any pricing or business model detail. The homepage meta descriptions confirmed the target sectors but were marketing copy only.
• ›Steps [11], [14], and [21] (FAQ pages for general, recruiters, and candidates) were the only substantive content-bearing fetches from Extracadabra's domain, yet the full HTML payloads were truncated in the response, suggesting the agent struggled to extract structured information from unstructured FAQ pages.
• ›Step [9] (jobs subdomain) confirmed a parallel brand surface but added no new business logic—it simply replicated the market positioning on a separate URL, indicating poor information architecture.
• ›Step [19] (CB Insights profile) and external web searches provided the financial data (€27M revenue, €15M 2025 funding, competitor list) that the agent ultimately cited—none of this came from extracadabra.com's own published content. The site does not publish its own business model, pricing tiers, or competitive positioning.
• ›The agent correctly noted that no dedicated pricing page exists (/pricing, /tarifs, /pour-les-recruteurs all returned 404s), forcing reliance on third-party business profiles to infer the commission-based model.
• ›The site is primarily French-language with no English alternative surface, limiting agent accessibility and cross-referencing capability.