The agent successfully understood Selecto's core business and positioning from the homepage and AI agents subpage, but encountered significant gaps: pricing is opaque (no public rates), several expected URL patterns returned 404s, and competitive differentiation claims lack supporting evidence. The site is navigable but incomplete for self-service evaluation.

• ›The homepage (steps [1], [5]) and dedicated AI agents page (step [10]) were the only successful content-bearing fetches; all guessed URLs (/hinnoittelu, /palvelut, /viitteet) returned 404s, indicating poor discoverability of service details or pricing.
• ›Pricing is explicitly non-public: step [10] states 'No pricing information is provided on this page' and directs to sales contact, yet the homepage claims it is 'listed openly on the price list'—a contradiction the agent had to flag as confusing.
• ›The agent correctly identified that 'Easy' product details (Easyhost, Easylead) were mentioned but not priced or detailed, requiring inference from model memory that this was a gap rather than content the site failed to surface.
• ›AI agent technical differentiation (data control, platform choices: OpenClaw, Hermes, n8n, OpenAI, Anthropic) came entirely from step [10]'s dedicated subpage—the homepage did not drill into these specifics, forcing a multi-fetch strategy.
• ›Web search (steps [9], [11]) returned results for unrelated services (SelectFI dealership software) and competitor agencies but no third-party validation of Selecto's claims, limiting the agent's ability to assess differentiation beyond what Selecto published.