The agent successfully assembled a comprehensive explanation of Nomba by relying almost entirely on web search results and prior knowledge, since the website itself is a single-page application that returns the same HTML skeleton for all routes. The agent correctly identified Nomba as a Nigerian fintech platform serving businesses and diaspora remitters with multi-currency and cross-border capabilities, but had to work around a poorly navigable website to piece together pricing and feature details from third-party sources and scattered documentation.

• ›The website is a client-side rendered SPA that serves identical HTML to every URL path (/pricing, /products, /checkout, /about, etc.), making direct page fetches uninformative. The agent could not extract content from nomba.com itself and instead relied on web search results [7, 13, 14, 16, 22, 25] to answer all substantive questions about pricing, products, positioning, and competitors.
• ›Pricing information came entirely from third-party sources (Swiftbills, Nairaland, Siliconafrica) and blog posts referenced in search results—Nomba's own site does not publish a unified, machine-readable pricing page. The agent's final answer cites external pricing sources, not nomba.com, indicating the site failed to surface this critical information directly.
• ›Product differentiation and competitive positioning had to be inferred from tech news articles and Nomba's own blog (discovered via search), not from any dedicated comparison or features page on the main site. The agent correctly flagged that feature parity vs. competitors and detailed checkout fees remain unclear despite multiple fetch attempts.
• ›The site's single-page architecture severely limited agent navigability—no amount of URL guessing or navigation would yield structured content. The agent had to shift almost entirely to web search (30%) and prior knowledge (20%) rather than relying on the website itself, indicating poor machine-readability and accessibility for agent-driven discovery.