The agent successfully gathered enough information to explain Certus AI's core offering, target audience, and differentiation, but encountered significant friction around pricing transparency. The homepage [1] and 'How It Works' page [5] clearly stated what the product does; the demo page [9] and internal comparison blog posts [13, 14] provided context on positioning versus competitors. However, Certus AI's website does not publish pricing publicly—the agent had to rely on external sources and prior knowledge to estimate the range, and explicitly called this out as a major gap. The site is moderately navigable for discovery but deliberately obscures commercial terms.

• ›The homepage [1] and /how-it-works [5] pages directly expose what Certus AI does (voice agent for restaurants, order-taking, reservation booking, payment processing, POS integration, 24/7 availability, bilingual support). This core positioning is clear and machine-readable.
• ›Pricing is intentionally hidden behind a demo wall. The /pricing and /features routes return 404 errors [2, 3], and the homepage [1] contains no price information. The schema.org metadata lists price as '$0', suggesting it's a placeholder. The agent was forced to infer pricing from third-party comparisons and prior knowledge rather than primary sources.
• ›The site surfaces differentiators through blog posts—specifically comparative analyses against competitors (Loman AI [13], Kea AI [14]). These pages were discoverable and detailed, but they required the agent to follow internal links rather than encountering a consolidated feature/comparison page.
• ›The /book-a-demo page [9] exists and is accessible, but there is no alternative conversion path (no contact form, no self-serve signup, no tiered pricing to browse). The site funnels all commercial inquiry toward a sales call.
• ›Multiple standard navigation paths (404s on /pricing, /features, /solutions, /contact, /use-cases) indicate a deliberately minimal site structure—only homepage, how-it-works, demo booking, and blog posts are live. This is typical of high-touch sales models but makes self-service evaluation harder for agents.
• ›The agent had to use web search [11, 12, 16, 17] to find external pricing estimates and competitor comparisons, as the site itself does not publish this information. This required 18% web search sourcing and reliance on 45% previous artifact/prior knowledge to assemble a complete answer.