None of this was supposed to happen. India’s fintech story was meant to be a triumph. A tale of leapfrogging legacy systems, of using data and design to make money accessible, fast, and friendly. And for a while, it was just that. UPI transformed how India paid. Everyone could send and receive money instantly. Into that stream came fintech startups, moving fast, raising big, and breaking old models. They made credit frictionless. The interface was intuitive. The promises were seductive. Small-ticket loans were available in minutes, with no paperwork, no collateral, no questions asked.

But as the loans piled up, so did the complaints. Most of these were about the recovery process. About harassment. About shame. And about silence — because it wasn’t always clear who the borrower had actually taken the loan from.

At some point, the Reserve Bank of India stepped in. In August 2022, it published a fresh set of rules for digital lenders. Apps had to tell customers who the real lender was. If you weren’t a regulated NBFC or a bank, you couldn’t lend. If you were merely a technology player, you had to say so. And customers had to know whom they owed money to — the app, or the entity behind the app.

For the casual observer, this looked like yet another compliance headache. But for people in the business, it changed the rules of engagement. A senior banker, one of the sharpest minds in India, offered me a blunt reading of what happened next. “The RBI asked a simple question,” she said. “‘Are you a lender or a tech platform?’ Because if you’re lending, you fall under our lens. If you’re just tech, you can’t touch the money.”

This question sparked a turf war. The apps claimed they owned the customer . They collected the data, understood user behaviour, and made the credit decision. But the banks disagreed. “We did the KYC. We underwrote the loan. We’re the ones accountable if things go wrong,” they said.

What the RBI did, quietly, was expose a basic confusion: Who owns the customer?

The apps had built their models on the assumption that most Indians had no credit history, but plenty of payment history: power bills, mobile recharges, DTH subscriptions. Algorithms could read these signals. Predict intent. Extend credit. It made for great storytelling. Investors loved it. Bangalore was abuzz with pitch decks showing how data could unlock Bharat.

But the reality, as always, was messier. Small-ticket lending in India has never been scalable without human intervention. You can’t always replace judgement with code. Recovery requires not just reminders and nudges, but people on the ground who understand the difference between a dodger and someone genuinely struggling.

The trouble was, while UPI made repayments possible with a single tap, it didn’t guarantee intent. And the algorithms, for all their claims of “learning,” weren’t always transparent about what they were learning or how they were making those calls.

“It was inevitable that the RBI would intervene,” the banker said. “This wasn’t arbitrary. The credit tech playbook had gone too far. And now we’re all recalibrating.”

The recalibration came again in November 2023, when the RBI raised the “risk weights” on unsecured loans. This was technocrat-speak for telling banks: hold more capital if you’re lending without security. In plain terms, it was a warning: slow down. The growth in personal loans had started to look like a runaway train.

Startups didn’t take it lightly. Some complained about overregulation. Others argued that the RBI was reacting to noise, not data. But behind closed doors, even they admit, the space has become overheated.

That brings us back to the borrower. For them, none of this is theoretical. From the gig worker in Noida who borrows ₹5,000 to fix a broken phone to the housewife in Nashik taking a loan for her child’s school fees, both have no idea whether the app a lender or just a storefront. Both don’t read credit agreements. They just want to know: Who gave me this money? What do I owe? And what happens if I miss a payment?

These are not unreasonable questions. What the RBI is trying to do is restore a sense of clarity. It’s trying to ensure that the infrastructure of trust gets back to being robust.

This isn’t about stifling innovation. It’s about reminding everyone, especially the fastest runners in the room, that a system built on data is still accountable to the people who feed it. Because credit isn’t just code. It’s a promise. And if that promise fails, the cost isn’t a write-off in a spreadsheet. It’s a life, sometimes quite literally.