The agent assembled a partial but incomplete answer about Finominal by combining web search snippets and prior knowledge, since the site itself—built with JavaScript—was not machine-readable. The agent successfully identified Finominal as a portfolio analysis platform for professional investors using factor-based return attribution, but could not access pricing details, feature tiers, or account/integration information directly from the website, forcing reliance on external sources and educated guesses. The site's navigability for an AI agent was poor: the homepage redirects to /home, most standard URL paths (about, features, docs, product) return generic redirect boilerplate rather than content, and the pricing page renders only as JavaScript that the agent cannot execute.

• ›Steps [1], [4]–[7] were routers/redirects: the root and several guessed paths returned only boilerplate HTML with a JavaScript redirect (if (window.location.pathname === '/') { window.location.replace('/home'); }), providing no substantive content. Only [2] (pricing) and [3] (home) returned Next.js-rendered HTML with actual page structure, but the agent could not parse the dynamic content to extract pricing tiers or features.
• ›The agent's actual sources for the final answer were primarily web search results [9], [10], [12], [13], [17], [18], [19], which yielded snippets about the company (founder Nicolas Rabener, London HQ, 2017 founding) and positioning (factor analysis, alpha measurement) but not detailed pricing or feature specifications. The site itself did not publish accessible pricing tier details; search results referenced 'starting at $120/month' but did not confirm exact tiers.
• ›The site is very poorly agent-ready: (a) multiple URL paths are guesses that route back to home rather than render distinct pages, (b) the pricing page exists but is JavaScript-dependent and not text-extractable, (c) no sitemap or robots.txt-accessible content map was fetched, (d) no structured data (JSON-LD, microdata) about pricing or company info was discoverable. The agent had to infer company purpose and differentiation from search engine summaries rather than from finominal.com's own documents.
• ›The agent explicitly flagged three confusions grounded in missing site content: (1) pricing tiers and feature breakdowns 'aren't clearly documented in publicly available text', (2) free tier details are undefined, and (3) integration/platform support info is sparse. These gaps are site failures, not agent errors, but they prevented a complete answer.
• ›The agent used prior knowledge (67% of fetches sourced that way per metadata) to fill gaps left by the site, including assumptions about factor investing, alpha analysis, and how Finominal differs from Seeking Alpha. Without search augmentation, the site alone would have yielded almost no usable information.