The agent assembled a comprehensive explanation of Razorpay by visiting the homepage, product pages (payment gateway, subscriptions, invoices, payouts, RazorpayX), and supplementing with web search results for pricing and competitive comparisons. The site's main product pages were discoverable via direct URL guessing and prior knowledge, but pricing details and competitive positioning required external sources; the agent worked around incomplete on-site documentation by bridging with search results and cited third-party sources.

• ›The homepage [1] and product landing pages [6, 7, 9, 10, 11] contained high-level positioning and feature lists but lacked detailed pricing breakdowns and technical specs; the agent had to search externally for the 2% + 18% GST pricing model that anchors the answer.
• ›Several reasonable URL paths (/why-razorpay, /products, /features, /solutions) returned 404s [3, 4, 5, 8], indicating the site's navigation structure is not fully discoverable via URL pattern guessing—the agent relied on prior knowledge to directly guess working product URLs instead.
• ›Comparative positioning versus Stripe and 2Checkout came entirely from external search results [13, 14], not from the site itself; Razorpay does not publish competitive analysis on its own domain, forcing agents to assemble differentiation externally.
• ›The pricing page [2] and invoicing/payouts pages [9, 10] loaded but the HTML response was truncated in the transcript, limiting visibility into what structured data or pricing tables they contained; the agent cited an external blog post [15] rather than the native /pricing page as its pricing source.
• ›RazorpayX banking product [11] was successfully discovered, but the agent noted it could not find detailed pricing for payouts, payroll, and corporate cards separately—the site surfaces the product but not the pricing granularity needed for a complete answer.