Razorpay Uses DataRobot AI to Counter Fraudulent Orders

​One unique aspect of the Indian fintech unicorn Razorpay is its ability to implement AI into its merchant transactional solutions rapidly. The company enables secure, hassle-free point-of-sale transactions on the device, with foreign currencies readily accepted. Developers integrated the DataRobot AI Platform within Amazon Web Services (AWS), which drives developer productivity, as it speeds up building and evaluating app-driven models by up to 10 times.

AI safeguards Razorpay’s ability to protect merchants against undeliverable or fraudulent orders without incurring major losses. Before the DataRobot AI solution, such orders ran rampant, with unfulfilled order costs rapidly escalating. Razorpay data scientists convened and, within 12 hours, applied AI to understand the scope of the problem and patterns involved in fraudulent orders. They created a robust model, rolled it out the next day, and effectively blocked malicious orders.

The rollout did not impact a single legitimate order, restoring functionality to the system with no downtime. It helped avoid the possible erosion of convenience and trust. This DataRobot AI model remains in place and is highly effective in order monitoring and drift detection or in-production machine learning model performance degrading over time. The model also helps ensure regulatory compliance by operating with built-in fairness and bias mitigation capacities.

SP Investment Fund

Shop