India’s fintech boom is entering a different phase.

The first decade was largely about building digital payment rails and getting consumers and businesses onto them. The next could be about what gets built on top of those rails — from digital credit and embedded finance to AI-powered customer service and underwriting.

That shift is already visible in the numbers. The Unified Payments Interface (UPI) processed 23.68 billion transactions worth ₹29.8 lakh crore in July. But Sohini Rajola, Executive Director of Growth at the National Payments Corporation of India (NPCI), says those numbers should not be mistaken for saturation.

UPI has around 240 million daily active customers and roughly 500 million monthly active customers, Rajola said. That is still well below the number of people in India who have bank accounts, internet access and mobile phones.

“Though it seems ubiquitous to us sitting in this room, we’re only about 240 million-odd daily active customers, and about 500 million monthly active customers,” Rajola said in an exclusive interview with CNBC-TV18 at the Banking Transformation Summit.

Her point is important for the next stage of India’s fintech story: even after a decade of rapid adoption, there is still a large pool of consumers and businesses that has yet to become fully digital.

UPI still has room to grow

UPI’s scale can make it easy to assume that the payments revolution is largely complete. But Rajola’s numbers suggest otherwise.

The opportunity is not limited to smaller towns and rural India either. Rajola pointed to gaps even in large cities, including among women and people who do not yet use smartphones or digital payments.

That leaves room for UPI to expand both its user base and the range of transactions carried out through the platform.

The next challenge, however, is not just adoption. It is economics.

The sustainability question

UPI has been built at enormous scale, but maintaining and expanding a digital payments infrastructure of this size also comes at a cost. That is where the debate around merchant discount rate, or MDR, becomes important.

If merchants begin paying a fee on UPI transactions, the immediate concern is whether some of them will shift towards cards or other payment methods.

The answer may depend on what merchants are already paying for payments.

PayU CEO Anirban Mukherjee pointed to cash-on-delivery as an example. For some large e-commerce businesses, COD can cost around 400 basis points once higher returns, fraud and other expenses are taken into account.

That makes the economics of a UPI charge less straightforward than simply comparing a UPI fee with a zero-cost payment option.

“Some amount of revenues in the system that help make up for the charges, that help make up for the costs for significantly more investments, I think it goes around in everybody’s case,” Mukherjee said.

He estimates that India is only about 30%-35% digitised when all forms of payments are considered, with particularly significant headroom in areas such as business-to-business payments.

Cashfree Payments Co-Founder and CEO Akash Sinha also expects online merchants to be able to absorb some additional payment costs. Businesses already factor in card payments, which are typically more expensive than UPI.

That does not mean MDR will have no impact. Smaller merchants and businesses operating on thinner margins could be more sensitive to any additional cost. But for larger online businesses, payment costs are already part of the cost of doing business.

The larger question is whether a sustainable revenue model can be created without undermining the affordability and interoperability that helped UPI reach its current scale.

Credit could be the next layer on UPI

If the first phase of UPI was about moving money, the next opportunity could be using the same infrastructure to move credit.

Credit on UPI has taken longer to gain traction than NPCI initially expected. But Rajola said the pace is now changing.

In the entire 2025-26 financial year, transactions involving customers using a linked credit instrument on UPI totalled around ₹1,000 crore. That figure is now about ₹500 crore a month, and Rajola expects it to rise further over the next two to three months.

The base is still small compared with overall UPI volumes. But that is also what makes the growth potential significant.

A customer can link a credit instrument to UPI and use it at the point of sale instead of relying on money sitting in a bank account. For lenders, the benefit is better visibility into where credit is being used. For customers, it can make access to and repayment of credit more seamless.

Rajola believes the market should eventually be doing tens of thousands of crores of such transactions every month.

Government-backed schemes such as Kisan Credit Card and Mudra loans are also being brought into the digital ecosystem, while a number of public sector banks are adopting the model.

The other side of the equation is repayment.

Kissht Chairman and CEO Ranvir Singh said UPI AutoPay has already changed the way digital lenders collect repayments, with the traditional NACH mechanism losing ground.

“The swing of the pendulum has completely gone to the side of UPI AutoPay,” Singh said.

For lenders, automated repayments can reduce collection friction and potentially lower recovery costs. For customers, it means fewer steps between taking a loan and servicing it.

That makes UPI more than a payments rail. It can increasingly become part of the infrastructure through which credit is distributed, used and repaid.

AI is moving deeper into fintech

The other major change is happening inside fintech businesses themselves.

AI is already being used for customer service and product development, but its bigger impact could come from how financial companies assess risk, personalise products and serve smaller businesses.

At CRED, Interim CEO Miten Sampat said customer satisfaction scores for support have increased by 15%-20% with the use of AI. The company is also able to do three to four times more with AI than it could previously, he said.

That includes faster product development and more intuitive customer support.

Financial services are particularly suited to this because customers often have complicated questions that can be expressed in many different ways. CRED is using AI to help users understand their credit scores and what they can do next.

The more interesting implication is what this could mean for credit itself.

“One of the core beliefs that we have at CRED is that your track record is an asset,” Sampat said.

A customer who consistently pays bills on time may be more creditworthy than someone with a limited credit history, even if traditional systems do not always capture that distinction fully.

AI could allow lenders to process a much wider range of signals and use them to make credit more personalised.

AI could change the way lending decisions are made

For digital lenders, the potential goes beyond automating existing processes.

Singh said traditional lending has largely focused on one question: whether a customer will repay. AI can allow lenders to look at a much broader set of variables — including the circumstances in which a customer is likely to pay, early signs of deterioration and behavioural patterns across millions of borrowers.

That could make underwriting more predictive and allow lenders to intervene earlier.

But more data does not automatically mean better decisions.

“One should not be suspicious of false precision which an AI can give, because billions of data points can produce nonsense in a very magnificent manner,” Singh said.

That is an important distinction for fintech lenders. AI can improve the speed and depth of underwriting, but it does not eliminate the need for sound credit judgement.

Smaller businesses could be the big AI opportunity

Cashfree’s Sinha sees AI having an equally significant impact outside lending.

Small and medium-sized businesses often lack the people and resources to run multiple processes that larger enterprises take for granted. Payments companies can increasingly use AI agents to perform some of those tasks.

Cashfree, for instance, is working on an agent-based system that could follow up with customers who abandon their shopping carts.

Traditionally, a business might have needed call-centre employees to contact such customers, understand why a payment failed or why they did not complete a purchase. AI can potentially handle much of that interaction.

The result could be a levelling of the playing field between large enterprises and smaller businesses.

“That’s what we believe is the true outcome of using AI in any of the processes. It’s helping you do more with fewer resources,” Sinha said.

Payments, data and credit could converge

This is where payments, AI and embedded finance begin to come together.

PayU’s Mukherjee believes AI could dramatically reduce the cost of software, making sophisticated digital tools available to micro and small businesses that previously could not afford them.

As those businesses digitise, they generate more structured payment data. That data can then potentially be combined with other structured and unstructured information to assess their creditworthiness.

In other words, payments could become the entry point for financial services rather than the end product.

A small business could use a payments platform, receive working capital based on its transaction history, manage collections through digital payments and increasingly use AI-powered tools to run parts of its business.

That is the essence of embedded finance: financial services become part of the platforms and workflows businesses and consumers already use, rather than something they have to seek out separately.

The opportunity is potentially much larger than monetising individual UPI transactions.

India’s fintech story is still in its early stages

For all the talk about India’s fintech boom, Cashfree’s Sinha believes the sector remains relatively small compared with the size of the country.

“We’re still in the early days of fintech,” he said.

That may sound surprising after UPI’s extraordinary growth, but the distinction is between the scale of India’s digital infrastructure and the extent to which that infrastructure is being used for broader financial services.

Millions of consumers still have room to move deeper into digital payments. Credit on UPI is growing from a very small base. AI is only beginning to change underwriting and customer service. And millions of small businesses have yet to fully digitise.

The industry’s challenge now is to turn that infrastructure into a broader financial ecosystem.

CRED’s Sampat believes India should eventually have three or four fintech companies among the world’s top 10, just as the country aims to build globally significant banks and technology companies.

The first decade of Indian fintech was about building the rails. The next could be about making more economic activity run on them — from payments and credit to business software and financial services.

UPI may already be one of the world’s largest digital payment systems. But if Rajola’s numbers are any indication, India has not yet reached the end of the payments story. The bigger opportunity may be what the country builds on top of UPI.

Watch accompanying video for full conversation.