AI Learning Recommendations for Faster Course Discovery vs Razorpay

Razorpay's feed is dominated by cross-border collection SEO — how Indian exporters receive EUR, AED, SAR and CAD — alongside playbooks on BBPS, e-NACH and multi-gateway routing. Against that backdrop one entry is a different kind of post: the launch of Vulcan, an in-house payments foundation model the company says is already running in production. It handles route scoring, cross-merchant fraud detection and checkout personalisation from a single model rather than the three siloed ML systems Razorpay describes running before.

The content and the product are approaching the same problem from opposite ends. The SEO cluster explains how to work around India's fragmented rails one corridor at a time; Vulcan is an attempt to make the choice of rail automatic. Razorpay is framing its own transaction volume — roughly 4 billion customer-to-merchant payments a year — as the asset that makes the model possible, which shifts its differentiation from gateway coverage to decision quality.

Razorpay states the goal is for every payment decision on its platform to run through the model, so the near-term move is extending Vulcan past routing, fraud and offers into the remaining decision points. Expect merchant-facing success-rate numbers to follow, since the claim only converts if it can be compared against an existing stack.