Roy: charting a path to achieve scale 
Corporate Report

L&T Finance focuses on retail

For L&T Finance, Lakshya pays rich dividends

Lancelot Joseph

Sudipta Roy joined L&T Finance Ltd. (LTF) in July 2023 as the Chief Operating Officer (COO) from ICICI Bank, where he was Group Head, managing diverse sets of businesses like unsecured assets, cards, payment solutions, student ecosystems, e-commerce and merchant ecosystems, millennial banking, and API banking. Roy’s appointment as COO and subsequent elevation as MD & CEO was part of the NBFC’s succession planning once the then MD & CEO, Dinanath Dubhashi, would superannuate from the company in April 2024.

Roy hit the ground running as he took charge as MD & CEO of LTF on 24 January 2024. “The overall organisation in terms of the philosophy towards business, technology, etc, has been in play for at least 12-14 months in terms of the go-forward strategy, for example, the Lakshya strategy, which is part of the overall L&T group objectives. Pre-Covid, LTF was largely a wholesale plus rural NBFC. Post Covid, around 2022, LTF started working towards slowly getting out of wholesale lending and fully focused on retail. That retail focus started as part of Lakshya, where our target was a CAGR of above 25 per cent,” explains Roy. 

Surpassing targets

LTF had targeted to achieve retailisation of 80 per cent, and retail return on assets (RoA) of 2.8 to 3 per cent. This is what it set the goal for FY26, ie March 2026. “As of the last quarter, we hit all those numbers. So, for example, if you see, the original target for retailisation was 80 per cent. Now, we have 96 per cent retail as of the last quarter ended 30 September 2024. The second thing we had targeted was greater than 25 per cent CAGR. Roughly, we have been doing 28 per cent plus year-on-year growth rate,” discloses Roy.

HSBC Global Research stated that LTF has the ability to bounce back sharply once operating conditions in MFI improve – they expect to see this improvement from December 2024. Such an improvement would lead to an uplift in disbursements growth, net interest margins (NIMs) plus fees, and credit cost estimates. At the moment, they do not anticipate the MFI stress to last longer than one to two quarters. “The wholesale business has shrunk; we now will move to a consolidated RoA of between 2.8 and 3 per cent by the end of FY26. Last quarter, we were at about 2.60 per cent. So, in a way, Lakshya 2026 targets were already achieved. So, we said, let’s define some things, which are starting at the leadership level and down to the ground level. All aspects might not be relevant for every cohort of employees, but some might be relevant,” adds Roy.

For the quarter ended 30 September 2024, LTF recorded a PAT of Rs696 crore (consolidated), up 17 per cent YoY, while the retail book stood at Rs88,975 crore, reflecting a YoY growth of 28 per cent. LTF has trimmed its marketing expenses and its HR recruitment in line with product launches, contributing to an improvement in operational efficiencies ahead. LTF aims to increase its retail share from the current 96 to 98 per cent by FY26 while keeping its wholesale business steady, Elara Securities India stated in its latest report.

LTF had defined five pillars of execution, which were rolled out in October 2023. What are the five pillars? The first of the pillars is to increase customer acquisition. The second pillar is to improve credit delivery and credit quality. To improve the quality of credit underwriting, the company wants to build world class credit infrastructure. The third pillar is to build world class digital infrastructure. The company wants to transform to a FinTech@Scale. The fourth pillar is to improve brand visibility. “One of the things that I realised in one of my discussions with SN  Subrahmanyan, Chairman of the L&T Group – I told him, as an outsider, I do not see LTF to be that visible in the market. So, that’s one of the things that we have to fix if you really want to become a retail company, because at the end of the day, retail customer demand flows from brand recognition. The L&T brand was very well known and quite visible. The L&T Finance brand name probably was not that visible, especially in the retail segment. We were primarily wholesale, where there was no need for advertising. So, if we wanted to become a full-fledged retail player with an equal presence in rural and urban areas, you required that brand identity and that brand presence,” observed Roy adding that the fifth pillar was to build capability. “That means you hire the best-in-class people, and you make sure that you build a future-ready organisation in terms of people capability. So, these were the five pillars that we defined in October 2023 and from there we started execution around these pillars.”

Silicon Valley approach

LTF has appointed Kavita Jagtiani, formerly of Pidilite, as its new Chief Marketing Officer to strengthen the brand’s visibility as a retail NBFC. LTF also wanted to make sure that the organisation is staffed with people who have a very strong technological DNA. “One of my objectives was to see if we can get people with the Silicon Valley approach, as they are more product-led. We in India are more comfortable building services. But there in Silicon Valley, the approach is holistically towards product building, end-to-end product building. So, we wanted to do that. The aim was that we would first try building the technology to develop it in-house rather than buy it,” says Roy.

Then, Ramesh Aithal from Elastic, probably one of the best commercial AI search firms based out of the Valley, joined LTF as Chief Digital Officer. Debarag Banerjee, a Silicon Valley professional with a Stanford PhD, who was Head of Data Sciences for Agoda, and previously Chief Data Scientist of Jio, came on board. In the second rung, LTF brought in people who have deep expertise in the Fintech ecosystem in India.

“To make sure that our distribution administration is far more granular, we sort of mirrored the structure of a bank. Like in a bank, you have four strong regional business heads — North, South, East, and West. We also transformed the organisation from a horizontal structure to a metric structure. We have introduced regional heads. In terms of capability, and as an organisation that is forward-looking, our compliance structures are absolutely world-class and very strong,” says Roy.

The Reserve Bank of India has been stressing a lot on making sure that there is adherence to quality processes in customer service, transparency to customers is ensured, and there is a great focus on cybersecurity, etc. Here, LTF has put together a very strong advisory framework. Four advisors in functional areas, who joined recently, are: Goutam Sanyal, who took care of credit policy and risk for retail products in ICICI Bank and advises on Credit, Risk, and Capital framework; Sankarson Banerjee, who used to be the ex-CIO of RBL Bank and previously Deputy CTO of the National Stock Exchange, joined to advise on Digital and IT Infrastructure; Kazi Zaman, who was previously MD of Everstone Capital, advises on Investor Relations; and Ganesh Kumar, who was Executive Director of the Reserve Bank of India, joined to advise on Risk and Compliance, respectively.

“We believe that we are one of the first to build a three-dimensional credit risk engine in the country. Traditionally, lending used to be done based on a single dimension, which is the basis of credit bureau data. This is a traditional 1970s-1980s Western model, which I personally believe has outlived its usefulness. The world is moving at an increasingly rapid pace, especially in India, where the velocity of transactions is exceptionally high in every sphere,” adds Roy.

We want to remain focused as an NBFC and grow as an NBFC for the next couple of years. Frankly, the question of whether we want to become a bank, etc, does not arise right now because we are in the process of growth in our current avatar

The second aspect is liabilities data. Now, banks always had access to credit bureau data and liabilities data. For example, if it is a bank customer, and you are giving them an auto loan, home loan, personal loan, or even a credit card, one can actually go to the credit bureau and look at their credit bureau score and their liabilities, figure out the balance between income and debt, which typically leads to a richer underwriting outcome. And that is why traditionally banks had a 100 to 150 basis points credit risk cost advantage over NBFCs, because they had access to the liabilities data, at least for their own customers. Now, because of ‘Account Aggregators’, it has become a relatively level playing field and LTF has access to a good amount of data.

Recently, 100 million consents on bank accounts were processed through ‘Sahamati’, the Account Aggregator network. “Initially, when we started, we had difficulties with the success rates of Account Aggregator. But we have tweaked the system, and now we are able to pull Account Aggregators with a 90 per cent success rate. This means that for every 100 customers, we can retrieve bank account statements for 90 of them,” says Roy.

Trust signals

The third aspect LTF now relies on in this new generation of underwriting is called ‘trust signals’. What does Roy mean by trust signals? For example, there is a lot of payment data circulating through various payment gateways. If you have made 15 transactions through a particular payment gateway over the past year, that payment gateway gradually learns about the nature and volume of your transactions. Some customers perform 15, 16, or 17 UPI transactions a month, while others may make 60 transactions a month. Due to the frequency and size of these transactions, as well as their locations, one leaves behind a significant digital trail with each transaction. Fintech companies have figured out how to predict income or affluence based on the density and volume of the digital transactions. Some of them have monetised this data as what they call ‘credit input’ or ‘credit insight’.

This is not a credit score; it is a credit input or a credit insight. The liabilities data through an Account Aggregator provides insights into your liquidity, while trust signals indicate affluence, spending patterns, or the density of the spending. A person with a high density of spending typically has a significant amount of liquidity and is highly digitally native. Therefore, the assumption is that such a person should perform well from a collection’s perspective. “So, when we started building the three-dimensional model, we identified six to eight trust signals. We also use a concept called micro-geography. For example, we have divided the country into grids, and now we know the loss rate per grid – the traditional loss rate, product wise. If a particular customer is coming from one grid, from a population perspective we know whether that grid is risky or not. And that risk weight goes into the underwriting algorithm. We wanted to build a machine so that when a customer comes in, we are not only able to pool in data from the credit bureau, we are able to fire the Account Aggregator, and then the trust signals as well. The processed data that comes in gives us a far more distilled understanding of the customer than was previously possible,” adds Roy, who calls this engine ‘Project Cyclops’ named after the Greek mythological creature – a monster or a superhuman being with a single large eye. This is completely developed in-house.

‘Project Cyclops’ serves as the gatekeeper of credit risk. Nothing escapes its notice. ‘Project Cyclops’ performs incredibly well compared to traditional algorithms. But it is a long haul in the sense that LTF is scaling up slowly as it is a new credit engine, and it has to make sure they do not make any mistakes. The initial results indicate that ‘Project Cyclops’ is performing well. It is separating risk much better than the old traditional algorithms. “We will continue to monitor it for another 4-5 months of continuous operation to be absolutely convinced that this is something that is extremely viable and gives us a significant order of magnitude difference in terms of credit underwriting.”

‘Project Cyclops’ is an omni-credit and omni-product engine, which means that you can underwrite any products through it, whether they be auto loans, personal loans, mortgage, tractor; one can underwrite anything through it. For example, ‘Project Cyclops’ went live specifically for two-wheelers. Currently, LTF is using ‘Project Cyclops’ only for two-wheeler loans because it is a high-velocity business. For two- wheeler loans, LTF has 16 scorecards on ‘Project Cyclops’. Next, the tractor business will go live. For tractors, LTF is taking a slightly different approach. “We are working with fintechs to digitise APMC data because predicting rural income with some certainty is essential for tractors. For example, if a customer is involved in cotton farming, some of our ‘Project Cyclops’ scorecards can assess the average cotton crop failure rates and the average Minimum Support Price (MSP) for cotton in that area, district, or state over the past few years. This allows us to estimate income and the risk of crop failure with greater accuracy,” adds Roy.

Holistic approach

LTF’s approach is to make the underwriting extremely holistic. Its in-house technology and application teams have already started working on building an Automated Portfolio Management engine as well. This is also something that Roy wants to build because a very large area of BFSI is not yet automated. In terms of technology, he wants to push the needle so far ahead that he wants to build systems that are a couple of generations ahead of some of its other peers.

Lastly, there is, of course, a great focus across the entire organisation on acquiring new customers. At the same time last year, the average monthly mortgage loan originations were Rs632.03 crore; now, it is close to R945 crore per month. The average two-wheeler loan origination was about Rs606 crore at the same time last year, and now it stands at R798 crore.

Overall, over the last 16 months, LTF has aimed to make the organisation more focused on outcomes, technology, people, capability building, and, finally, on brand and product development. “For example, our first reengineered product launch was ‘The Complete Home Loan’ in which we reimagined our entire complete home loan product and launched it with a good 360-degree campaign during the IPL. We introduced Sky RM, which is a dedicated RM, and a Home Decor Finance product that is embedded with our home loan. It has been very well received in the market, and is performing quite well,” adds Roy, who is focused on reimagining all products and strategically pushing the brand for each individually.

So, even though on the whole, LTF has picked up speed, a lot of things remain a work in progress. At times there are environmental headwinds. For example, last year the tractor business was very sluggish due to paucity of rainfall. So, sustaining the tractor business required a lot of effort. “This year, we are hopeful that the good monsoon season will positively impact our business, as a large proportion of it is based in rural areas. As 47 per cent of our business is rural, 53 per cent of our business is urban right now,” says Roy.

The other thing that LTF is doing is focusing a lot on AI and building real life AI use cases. For LTF’s microloan customers, when the employee visits, he captures pictures of the dwelling, inside the dwelling as well as outside of the dwelling. And there is no human intervention. It is basically the algorithm that looks at the pictures.
“We have trained the algorithm to identify whether it is a Kacha house, whether it is a Pucca house. We have trained it to identify whether
there is an electricity connection inside the house or not and figure out whether there are white goods in the house, or not. So, it
will tell you all that, and then it will tell you the overall condition of the house and then based on that, it will formulate a standard of living index,” adds Roy.

“LTF has filed the patent for it. And this is also one of the objectives – that as an organisation, we want to now start creating intellectual property, especially around AI. So, all our senior management has now been trained in AI and the implementation of AI. For example, all our General Executive Council members, can now actually write prompts and then train the model themselves. We have done our senior management training on AI to that extent. We are committed to building these real-life AI use cases, and our teams are very focused on this goal. Over the next couple of quarters, you will see us introduce several powerful AI use cases. Another key area of focus is using AI in customer service, where we plan to implement AI solutions as well,” says Roy.

The objective is to make sure that LTF is seen as a place where cutting edge work is done in terms of credit analytics and credit models in terms of AI, customer service, UI/UX, and our digital as well as mobile banking. Basically, LTF’s customer-facing PLANET app has emerged as a powerful digital channel for customers and has crossed more than 1.25 crore downloads till September 2024, comprising more than 13.5 lakh downloads from rural customers. For LTF, a lot of customer service and payments happen through our Planet App right now.  

Growth plans

Finally, LTF’s plans and projections are guided by the Lakshya goals. “We have met most of the Lakshya goals, but one of the goals is that we have said 25 per cent CAGR in terms of AUM growth year after year. We will try to make sure that we hit that number. So, till FY26, that will be our guiding principle. There will be another plan after that, and we will decide the next course of action accordingly. However, we remain committed to 25
per cent year-on-year growth at least until FY26.”

Originally, one of LTF’s objectives was that they wanted to become a full-scale bank. “But because of various things that did not materialise. So, as of now, we are charting a path of becoming and achieving scale as a diversified NBFC, straddling both rural and retail. We are a very unique NBFC with 47 per cent of our book in rural areas and 53 per cent in urban areas. So, in a way, we are a very balanced organisation in terms of our customer profile. We want to remain focused as an NBFC and grow as an
NBFC for the next couple of years. Frankly, the question of whether we want to become a bank, etc, does not arise right now because we are in the process of growth in our current avatar. However, if the Reserve Bank of India at any point in time gives an indication that they are amenable to converting NBFCs into universal banks, etc, then it is for the
L&T Group to take that decision. So, from a strategic perspective, it is a bridge that we will cross at that time,” concludes Roy.