From FOMO to Function: Can AI transform India’s fintech and banking backbone

In AI parlance, greenfield refers to building entirely new systems or products designed natively with AI from the ground up, while brownfield refers to adding or integrating AI into existing legacy systems or products that were not originally built with AI in mind.

At GFF, the buzzword was AI,  but FOMO ran the floor

Nowhere was this paradox clearer than at the Global Fintech Fest (GFF) 2025, held at Mumbai’s Jio World Centre from October 7–9. The theme this year, “Empowering Finance for a Better World Powered by AI”, captured the mood perfectly.

From panels on agentic AI and small-language models to demos of multilingual bots, the three-day event was a showcase of what’s possible,  and what’s still aspirational. NVIDIA’s presence was particularly striking: in collaboration with NPCI it unveiled the “Bharat AI Experience Zone”, a dedicated arena to demonstrate real-world AI use cases in banking, payments, and credit.

Voice AI stole some limelight, DCB Bank’s new multilingual AudioBot built with Ringg AI, and Bengaluru-based Gnani.ai’s lifelike “AI-generated Digital Human” that could talk, blink, and emote in real time,  but even these innovations were more showcase than scale-ready.

Ganesh Goplan, CEO at Gnani.ai believes the requirements for enterprise AI will be very different from the personal use cases of ChatGPT.

“You can easily wait for 5 seconds for ChatGPT to respond to your queries, but the moment your bank app takes a few seconds extra to show your bank balance, you would slam the phone,” he says.

“There’s a lot of FOMO,” admitted one startup founder. “Last year, the RBI wasn’t very keen on AI. This year, the sentiment has reversed—seeing the global mood and the pressure to stay ahead, no one wants to miss this bus. Startups, even banks, are trying something, but it’s all very experimental right now.”

“Every boardroom is pushing their middle management to launch an AI product,” Gopalan concurred. He added that banks and enterprises have already started seeing good return on investments (RoI) and better business outcomes – that’s fuelling further interest.

That summed up the tone well: excitement mixed with realism. Everyone wanted to ride the AI wave, but most admitted they were still learning how.

The fear factor: Don’t mess with money or the regulator

“AI adoption in financial services isn’t about letting a thousand flowers bloom,” said Neetu Chitkara, Managing Director and Partner at Boston Consulting Group (BCG).

“We’ve seen that random bottom-up experimentation doesn’t work. You need a clear top-down blueprint—defining three or four core areas where AI will deliver measurable value. That strategy upfront determines whether a firm scales or stalls.”

Chitkara added that clean data, leadership alignment, and responsible AI policies are crucial. “If your data isn’t right, even an impressive PoC won’t sustain,” she said, warning that skipping ethical checks can push firms “two steps forward and five back.”

Fintech founders and investors alike agree: India’s fintech ecosystem cannot afford a reckless AI race. The stakes, trust, regulation, and risk, are too high.

As one neobank founder put it, “AI is across every segment. But it is still vanity. Everyone is in FOMO and going with the global theme. But so far it is still at a very nascent stage for early-stage and growth-stage companies. Bigger players are striking partnerships and are trying to do something meaningful. But this is going to take some time. However Agentic AI is getting used across. But India is a very top-down company, it should start from the regulator. Everyone is trying to decode the syllabus without the textbook.”

That “textbook,” as the RBI Innovation Hub (RBIH) CEO Sahil Kini explained, is the RBI’s own AI report,  a free, publicly available guide on how artificial intelligence should be deployed in financial services.

“Anyone building AI in fintech without reading that report is doing themselves a massive disservice,” he said, adding that it outlines seven core principles, including trust, people-first design, and explainability. “If you build without keeping those principles at the heart, you’re on shaky ground.”

He added that RBIH is already building AI-driven public infrastructures like Mule Hunter, which detects mule accounts, and the Digital Payments Intelligence Platform, which will assign a real-time risk score to every digital payment in India.

As the regulator’s benchmarking suggests, fintechs must tread carefully. The financial services industry thrives on stability and trust; a mis-step in AI implementation could undermine the entire structure.

Why AI strategy can’t be a “let 1000 flowers bloom” experiment

According to Chitkara, both banks and fintechs are at different stages of AI maturity. “Most are still in what we call the ‘deploy’ phase, using tools like ChatGPT for productivity or automating outbound calls and credit memos,” she explained.

“A few are moving toward the ‘reshape’ stage, rethinking entire functions like marketing or customer engagement with AI at the core. But very few globally have reached the ‘invent’ stage—where AI creates entirely new business models or products.”

What AI could actually fix

Beyond chatbots and fraud-alerts, AI’s most promising potential in fintech lies in efficiency, inclusion and trust.

BCG’s Chitkara believes that AI at present is mostly being used as a cost lever rather than a revenue driver. Most of the firms are using AI to get things done faster or with few people to bring down cost. But on the revenue side, using AI for smarter customer targeting, product cross-selling or new digital offerings, adoption is still limited. “Thats the next frontier,” she said.

“Right now AI is a mindset, however the faster companies pick up, the faster we will see more innovation,” said Bhavin Patel, founder of LenDenClub.

“We are on right track but we need to get right product integrations. It is not easy as well. On credit we are trying to do some self-learning models. We are trying to build one complete AI app, we have built within 26 days, final user testing phase, we will launch it soon. We see that the first use case AI will solve is reduce manual intervention and increase efficiency. Then comes Agentic AI, some call centre operations are done by that and some reporting we are using AI. This is mostly for data summarisation etc.”

Gnani.ai has been working with financial institutions on marketing, collection, better operations and even customer support. “For instance, we have helped a customer improve their collection rate by 8%, reduced their NPAs, in just the last three months we have been working with them,” Gopalan shared.

The first wave of meaningful adoption will likely be in the behind-the-scenes plumbing: credit underwriting, risk scoring, collections, compliance, voice bots handling vernacular queries rather than flashy ‘agentic’ interactions.

The big prize is personalisation at scale, smaller loans to underserved segments, vernacular voice interactions, smarter merchants, faster agentic bots and everything anchored in India’s payments-infra base.

Startups like Gnani.ai are powering this shift by offering AI-driven voice analytics and conversational intelligence platforms that help banks and insurers engage with customers in regional languages and manage large volumes of voice data more effectively.

Others such as Zeta, Yellow.ai, Haptik and E42.ai are integrating generative AI into customer-service and backend operations, automating communication, compliance checks and data processing.

These solutions not only enhance customer experience, but also help institutions meet the regulator’s increasing focus on responsible AI and data security.

Even payment technology giants such as Mastercard are innovating fast. Nitendra Rajput, SVP and Head, AI Garage at Mastercard shares how AI Garage offerings are solving for transactions.

“We are using AI to identify customers for banks who habitually have huge debt on their credit cards and then don’t pay up. We use AI to track merchants who are setting up fake ecommerce websites too,” he said.

In agentic commerce, Mastercard is developing AI bots and solutions that can narrow the customer’s search by quickly scanning through 3000 websites to find the best deal and variant of a product they are looking for. Manually, the customer could have only checked out three to four websites. The payments too are being automated using AI for such search-based purchases.

Investors smell an opportunity

While fintechs tread carefully, investors are already placing big bets. Z47 (formerly Matrix Partners India) is ramping up investments in tech-first fintech and AI startups, both in India and globally.

“A third of our investments right now are in global companies with an India edge,” said Vikram Vaidyanathan, Managing Director at Z47, noting that founders now operate globally from day zero.

“Firms like Oxyzo are adopting tech in customer outreach, underwriting, and loan workflows like never before,” he said.

In consumer lending, digital-first cards and UPI-based credit lines are boosting penetration at lower cost, with startups like OneCard and Jupiter leading the way.

“We invest in companies that marry AI journeys with digital journeys—AI-first payments, AI-first InsurTech, and consumer experiences,” Vaidyanathan added, citing Grey Labs and Ulka.

“The firm remains bullish on core financial services but prioritises digital experiences and trust-building. Earlier, trust came from a big bank branch. Today, phenomenal digital experiences build trust.”

TVS Capital Funds Chairman Gopal Srinivasan added: “There are some very significant changes happening. Like BharatGen is doing some work on legal front. We are also looking at some of these startups. AI is seen as a democratizing force in India, with investment opportunities in ‘picks and shovels’ rather than complex AI agentic platforms.”

Investor interest is real and rising, but they’re watching for true metrics, not just highflying demos.

Fintechs experiment, giants commit

While startups experiment, large players like Paytm and Razorpay are moving from talk to action.

Paytm founder Vijay Shekhar Sharma announced an AI-powered point-of-sale (PoS) device for merchants. “If I have a fashion store, I can ask, ‘How is my business going?’ or ‘Did I receive money from the bank?’ and get answers in my local language,” Sharma said.

The device will run a locally trained, smaller AI model, operating without internet. “Every founder needs one teammate or co-operator; this AI will be that.”

Razorpay, NPCI and OpenAI are piloting Agentic Payments on ChatGPT—allowing users to discover, compare and pay for products directly within the chatbot using UPI. The pilot is a bold sign of where fintech + AI is heading: transactions via conversational agents, embedded commerce via trusted infrastructure.

These efforts show that the marquee name fintechs have accepted that AI is not optional,  it’s integral to the next platform wave.

A slow climb, but an inevitable one

AI adoption across BFSI remains uneven—deep in fraud detection, voice analytics, and customer support, but shallow in underwriting and full-loan automation. Yet the momentum is irreversible. Fintechs, being more digital-native, are moving faster than banks, even as they realise the paradox: what they do today with 300 people could soon be done by 30 using AI.

As Ramki Gadipati, co-founder of Zeta, put it:

“We are security, compliance and cloud the three fortes we always talk about, but now with AI shaping up for the enterprise especially for fintechs and banks is a different ballgame in itself; comes with its own challenges. It is not just about trying a tool but deeply integrating in functions is different. We are doing a lot of researching, experimenting and toying a lot in this area.”

Payment gateways like Mastercard are especially doubling down on cybersecurity in the fintech domain in the AI era, apart from improving speeds of digital payments and reducing friction.

“Boom of AI is also helping bad guys get into finance easily using AI. Fintech has to be the first industry to adopt AI quickly given these scenarios,” Rajput said.

These sentiments mirror where fintech stood in 2015,  brimming with potential yet wary of over-stepping. AI won’t disrupt fintech overnight; instead, it will seep in through small, steady wins: a smarter underwriting engine here, a fraud detector there.

And perhaps that’s how it should be. As the RBIH chief said: “I do not think disruption is necessarily desirable in financial services. Some amount of stability and sensible progress is what the sector really needs. You want excitement in certain parts of your life, not around your bank balance. Predictability here is a good thing.”

While most sectors see AI as a sprint,  a race to avoid missing the bus , India’s fintech ecosystem views it as a marathon, one that demands patience, regulatory alignment, and long-term integration.