The agent successfully built a comprehensive explanation of Riot IQ's purpose, pricing, target audience, and competitive positioning by assembling information from the site's homepage, pricing page, features, about-us section, and comparison articles. However, the site's heavy JavaScript rendering and client-side architecture made detailed content difficult to extract from the raw HTML responses, forcing the agent to rely on prior knowledge and web search results (36% each) to fill gaps and validate findings. The agent identified several confusing or missing aspects (data privacy, governance structure, professional use cases) that were not clearly surfaced on the site itself.

• ›The homepage (step [1]) and pricing page (step [2]) returned valid 200 responses but rendered as truncated Next.js JavaScript bundles, making it impossible to extract actual pricing tiers, feature lists, or value propositions from the raw HTML. The agent had to infer structure from prior knowledge rather than parsing page content.
• ›Comparison articles (step [13]: 'Top 10 Online IQ Tests') and the about-us page (step [15]) were the most useful site-sourced resources, likely containing positioned messaging and team credentials, but again only URLs were accessible—not the actual rendered content that would explain differentiation vs. Wonderlic, Prepopedia, etc.
• ›The agent retrieved 8 pages from riotiq.com but only cited 6 URLs in its final sources. Steps [6] (FAQ) and [9] (blog) did not contribute to the final answer despite being fetched, suggesting they contained no novel information or were inaccessible due to client-side rendering.
• ›Web searches (steps [8], [12], [14], [16], [17]) returned snippet links to Riot IQ's own articles (e.g., comparison pieces) but Goodfirms and Crunchbase returned 403 blocks, limiting external validation. The agent filled pricing details and team credentials partly from search snippets and prior knowledge, not from the primary site fetches.
• ›Critical omissions the agent flagged (data privacy, professional use cases, enterprise/B2B options, governance) are likely documented on the live site but were not accessible through HTML parsing alone, indicating poor machine-readability of structured information.