Daybreak - AI has made its way into banks. But the gates aren’t wide open

Episode Date: May 20, 2025

For decades, the whole process of getting a loan approved was infamously painful and long winded. But now things have changed. Getting a loan is a whole lot faster than before. And that’s because of the disruptor to end all disruptors — artificial intelligence. A bunch of companies have entered the scene with specalised AI tools to speed up different aspects of the loan-approval process.  In fact,  Indian AI startups have managed to raise nearly 750 million USD in 2024 and the banking and financial sector was one of the top drivers of this growth. Now at first glance, it seems like a win-win for both the borrower and the bank. But there’s a catch. This surge has come with a lot of scrutiny from the RBI. Tune in. *This episode was first published on Jan 15, 2025Tell us what you thought of this episode. You can text us your feedback on WhatsApp at [phone]Daybreak is produced from the newsroom of The Ken, India’s first subscriber-only business news platform. Subscribe for more exclusive, deeply-reported, and analytical business stories.

Hi, this is Rohan Dharma Kumar. If you've heard any of the Ken's podcasts, you've probably heard me, my interruptions, my analogies, and my contrarian takes on most topics. And you might rightly be wondering why am I interrupting this episode too. It's for a special announcement. For the last few months, I and Sita Ramon Ganeshan, my colleague and the Ken's deputy editor, have been working on an ambitious new podcast. It's called Intermission.

We want to tell the secret sauce stories of India's greatest companies. Stories of how they were born, how they fought to survive, how they build their organizations and culture, how they manage to innovate and thrive over decades, and most importantly, how they're poised today. To do that, Sita and I have been reading books, poring over reports, going through financial statements, digging up archives, and talking to dozens of people. And if that wasn't enough, we also decided to throw in video into the mix. Yes, you heard that right. Intermission has also had to find its footing in the world of multi-camera shoots in professional studios, laborious editing, and extensive post-production.

Sita and I are still reeling from the intensity of our first studio recording. Intermission launches on March 23rd. To get alert, as soon as we release our first studio recording, episode, please follow intermission on Spotify and Apple Podcast or subscribe to the Ken's YouTube channel. You can find all of the links at the ken.com slash I am. With that, back to your episode. Back when Disha and Abinaw Sharma applied for a home loan, they had to jump through all sorts of hoops before their application was eventually approved. Between all the paperwork, the agent visits, zeroing in on the payment and interest rate, the property assessment by the bank

The Indorpese couple waited an entire month before their loan was approved. And this is not a one-off example. The whole process of getting a loan here in India, and by that I mean any kind of loan has been infamously painful and long-winded. And that's mostly because of all the steps involved, each of which will typically be carried out by a bunch of different people at different levels of the banking and financial sector. But today, I come to you with both good news and bad news.

The good news is that the whole process of getting your loan approved is now a whole lot faster than before. And that's because of the disruptor to end all disruptors, AI or artificial intelligence, of course. You see, a bunch of companies have entered the scene with specialized AI tools to speed up different aspects of the loan approval process. So, just take Bangal-based Navank for instance. This two-year-old fintech counts players like Muthut finance, Vastu-Hawar, housing finance and APAC financial services among its clientele. Its claim to fame is an AI tool that speeds up property evaluations, which happens to be the slowest part of the process.

Usually financial institutions need approval from an impaneled service provider who will visit the property and then either give a red flag or a green flag. And then this process is repeated over and over and over again. Now, with AI tools, the credit manager simply has to take photographs of the property. The AI tool analyzes these images and provides updates on the progress in construction. And voila. Based on this data,

bankers can assess property quality and make more objective data-driven decisions. This is just one example. Now, at first glance, it seems like a win-win for both the borrower and the bank. This is particularly helpful in smaller cities and semi-urban areas where

housing projects often lack proper legal paperwork and don't have bank representatives stationed at developers' offices. But there is a catch. And that brings me to the bad news. So Indian AI startups have managed to raise nearly $750 million US dollars in 2024. And banking and financial sector was one of the top drivers of this growth. But the surge has also come with a lot of scrutiny from the Reserve Bank of India. In fact, it even set up a committee to evaluate the usage of AI in banking and financial services. It's called Free AI.

That was the Ken reporter Akriti Bhala. Like she said, the RBI has some very strong news on generative AI. It's hyper aware of the fact that with all the convenience comes a fair share of biases and errors. Welcome to Daybreak, a business podcast. from the Ken. I'm your host Rahal Philippos and I'll be joining my colleagues Nikda Sharma every day of the week to bring you one business story that is worth understanding and worth your time.

Today is Wednesday. Hello, dear listeners. Did you hear the news last week about Amazon's prime video introducing ads on their platform next month onwards? It looks like India's OTT game is entering a whole new era. Ads are in, subscriptions are evolving. But what is really going on behind the scene?

And what does this mean for us, the viewers, and what does it mean for the world of advertising? This week, Swati Mohan, the ex-head of marketing at Netflix India, will be joining Rahil and I for a chat on daybreak. Swati is someone who has had a front row seat to the rise and the evolution of India's streaming giants. From this shift to an AVOD-S-V-O-D hybrid model to regional audiences rewriting the rules for OTT, this conversation is going to be packed with hot takes and sharp insights and some seriously good stories. So watch out for the episode. And now back to Rahil. Let's talk about how a bank typically decides whether a particular customer is eligible for credit.

Generally, it all comes down to the Sibyl score. Now, this is essentially a summary of your entire credit history. Based on this three-digit score, a bank can assess whether someone qualifies. for a loan and how much they can borrow. But the thing is, generally, the Sibyl score alone does not capture the full picture. It's easy to, quote, unquote, hack the system. For instance, an applicant may have a great Sibyl score because they've taken and repaid a bunch of smaller loans.

But that does not necessarily mean that they can afford a larger loan. Their Sible score may say otherwise. So it doesn't always reveal their true repayment capability. Now, that's exactly where Gen AI comes in. It could potentially help banks analyze spending and saving patterns, income levels, and much more. Basically, banks will be able to receive more granular data on loan applicants, things like their spending patterns and behaviors. Essentially, they'll have access to a broader set of data to assess creditworthiness,

which will make the whole process a lot fairer. Just take Bangalore-based startup bureau for. instance. It offers tools to crunch, analyze and visualize customer data, including behavioral patterns. These tools help banks, non-banks and fintechs determine when they should offer loans. But these AI tools can help with a lot more than just underwriting. They can also help with other functions. I'm sure you're aware of how sales teams at different financial institutions are usually going all out to meet targets towards the end of each quarter. Now, generally, it is the compliance team that bears the brunt of their enthusiasm.

Because of the volume of applications, they generally perform basic hygiene checks, verifying addresses, Adhahar Info, and Sybil's cause, which ends up leaving scope for a lot of incomplete or fraudulent applications to slip through the cracks. Now, that is where Bangalore-based Neurofin AI comes to the rescue. You see, tools like Neurofin's help spot errors related to legal and policy regulations that a human compliance officer may miss. The company's founder, Vijay Makhijiani, told the Ken that it acts as a gatekeeper for compliance. It also processes loan approval documents in one-tenth the time and for half the cost.

One of its functions is to flag applications where agricultural land is provided as collateral, which isn't allowed under the RVI's rules. But all things considered, Gen AI startups can improve how we are. assess and monitor portfolio risks by providing better early warning systems. And that's important, considering that currently the systems in place at banks and other financial institutions generally flag risks too late. AI-enabled tools can help detect potential risks and areas of non-compliance on a real-time basis, making it possible to contain them before real damage is done.

It isn't just helping with detecting risk, is also helping with internal compliance. These tools help banks stay on top of regulatory requirements issued by bodies like the RBI and the Sebi. For instance, on Finance AI, a startup backed by Silver Needle Ventures and Indian Angel Network, developed a tool called Neo-GPT Compliance OS agent, which helps compliance officers track regulatory updates across different departments. Depending on the type of financial institution, whether a bank, non-bank, or a mutual fund, Neo-GPT creates a tailored compliance checklist. Whenever a new regulation or a circular is issued by a regulatory body,

the tool scans it, identifies a necessary action for that particular financial institution, and then generates a checklist. So a bank's checklist would be different from that of a mutual fund. But in the end, it comes down to the compliance officer to make sure that the required disclosures and reports are accurately submitted to the regulator. So it's not like human intervention is ruled out completely. Generally, banks and other financial institutions can take up to a year to wet generative AI tools before implementing them. But most of them seem convinced of the value they offer.

Bajajaj finance, for instance, intends to scale the use of Gen. AI across various loan categories. I know right now it may seem almost too good to be true. But don't be fooled. There's more to these AI tools than meets the eye. More on that in the next segment. Before we jump back in, I just wanted to share some exciting news with you. Daybreak is growing and we are looking to grow our team too. Now, if you're a journalist who listens to Daybreak and thinks,

I wish they'd cover this or I know exactly what would make Daybreak even better, we would love to hear from you. We're looking for somebody who's curious, who's driven and passionate about business and tech news. If that sounds like you or somebody you know, please check out the details in the show notes. Applications are open now and with that, let's get back to the episode. The thing is, algorithms like these aren't free from biases. And that's something the banking system just cannot dodge. The Ken spoke to three deep tech experts who said that generally these tools recycle the same biases baked into the system by humans.

So, for instance, reports recently pointed out that lenders were far more likely, to reject the loan applications of female entrepreneurs. Even when their socio-economic backgrounds are identical to their male counterparts, their loan applications are often delayed or turned down. Then there's also an income bias. 60% of middle-income households are credit underserved. Now, in this kind of situation, the stakes are very high because biased algorithms could lead to erroneous decisions.

Or worse still, they can reinforce gender, racial, and other forms of discrimination. And this isn't just an India issue. Even in the Western world, in places like North America and countries across Europe, AI has been called out for discriminating against black and Hispanic communities. But FinTech firms working in the space argue that even with biases in the system, generative AI can be trained to correct them. Loan applications may have been rejected in the past because of the applicant's home address,

gender, income or socioeconomic background, the evaluation system would be triggered by any of these traits and tag the applicant as risky. With generative AI, however, the evaluation is more nuanced. Anil Tadimeti, the Director of Strategy and Regulatory Affairs at Bureau, said in this case, an applicant's own behavioural history is taken into account, not just the collective credit history of any cohort that they belong to. Plus, it can also help detect friends.

But still, the RBI isn't fully on board as yet. You see, while fintech founders and investors are betting on Gen.EI to disrupt banking, bankers and regulators are still treading quite carefully. The thing is, as things stand, there are no regulations that govern the use of AI and financial services. And therefore, there is a lot of risk involved in automating an industry that holds significant sway over the economy. Then there's also the issue of data security.

Now, as per the RBI's regulations, non-banks are not allowed to outsource core functions like audits or compliance decisions to third parties. So while AI tools can help with decision-making, the final say has to lie with compliance officers. At the moment, generative AI adoption in banking is mainly limited to customer service functions, like those handled by chatbots or the grunt work of property evaluations. And so far, they've delivered decent results. But there is a way to go before they are ready to take on complex or risk-heavy assets. Until then, people like the Sharmas and Indoor might be stuck waiting for months or dealing with pin-code bias that residents of tier two cities endure during loan applications. Daybreak is produced from the newsroom of the Ken, India's first subscriber-focused business news platform.

What you're listening to is just a small. small sample of our subscriber-only offerings. A full subscription unlocks daily long-form feature stories, newsletters and podcast extras. Head to the ken.com and click on the red subscribe button on the top of the website. Today's episode was hosted by Rahil Filippos and edited by Rajiv Sien.