Harshil Mathur: AI Is Compressing Every Moat

Mathur began as a coder and found the payments problem while trying to accept money for a side project in India. Razorpay first targeted educational institutions, but colleges had little reason to make fee collection easier. Startups cared about the problem, so the company pivoted to serve them.

Razorpay waited nearly a year after YC to process its first live transaction while it secured approvals and certifications. Mathur saw that difficult entry requirements could become a moat because competitors faced the same process. When its bank partner shut down service for about 50 merchants, the team called each customer, explained the situation, and restored service with another partner within four or five days.

Mathur said Razorpay kept spending lean because businesses paid for concrete value, rather than engagement metrics. In 2016, while larger gateways waited to see whether UPI would succeed, Razorpay became the first payment gateway in India to go live with it. After demonetization, major merchants including Zomato, Swiggy, and BookMyShow joined, giving Razorpay an early advantage.

Mathur said AI is making software faster to build, so companies need to decide sooner what to build and move before the market shifts. Razorpay redesigned its platform around integration, onboarding, and support, and redirected teams toward the new approach. He also warned against founders retreating into manager mode: they should stay involved in product vision and other core decisions while delegating the rest.

Mathur advised founders to look beyond how easily AI lets them build and launch software. The key question is whether they care enough about a problem to work on it for the next 10 years. He said sustained effort on a problem founders deeply understand is still central to building a company.