L&T Finance launches 'Project Cyclops', a three filter mechanism to test credit quality of borrowers: Sudipta Roy

Apart from this, he added that the company has drawn $50 million of the total $125 million raised from the Japan International Cooperative Agency to finance women borrowers.

Edited excerpts: 

There is talk of a new initiative, can you share details?

The beta version is already live. The company has deployed the beta version of ‘Project Cyclops’, a cutting-edge credit risk assessment and automated decision-making digital credit engine which harnesses the power of Artificial Intelligence (AI) and Machine Learning (ML) to determine the repayment capability and credit quality of potential customers.

In September last year, we thought we will build an integrated credit engine which we call the three-dimensional credit engine.

This word is very important because the first dimension is the bureau, the second is liabilities data and the third dimension is trust signals. They are actually statistically validated data regarding income levels, consumption levels obtained either from the UPI transaction flow data in many contexts or micro geography which is geolocation based intelligence and this is the third dimension or the trust signal dimensions.

If the customer is a thick bureau, we will probably not need to do anything else.

If the customer is a thin bureau we need to do a liabilities pull as well as as a trust signal pull and if the customer is completely new to credit that means the customer does not have a bureau track or  a thick liability stretch. Then you actually lean on trust signals as well and try to see if there is data that you can pick up for that particular customer.

We have built this entire engine over the last six to eight months, which we go by the sort of name Project Cyclops.

Also read: PhonePe expands loan offerings with launch of secured lending

Why is this system important?

The Project Cyclops engine, though we probably cannot say with a hundred percent guarantee, it but it is probably one of the very first sort of three dimensional integrated credit engines in the country that does bureau, account aggregators/liabilities and trust signals in one pass, and with very low latencies in a very fast manner and it is also an omni-product that means you can do any product on it.

Who has built this and how much cost was incurred?

It has been built up by our in-house cross-functional team, the data science team or technology team. Our credit teams,  with a 100 member team worked continuously on it for the first six months.

If I were to put a number to the people cost, maybe it has cost us, less than Rs 5 crore to build this system.

What external help have you taken?

One fintech, PayU for example, its score is quite well developed. So PayU has helped us with the PayU scores on the account aggregator side. We are working with many account aggregators, primarily Parfios. We are working with fintechs on the Geo IQ side, for micro geography data.

At how many locations is the Beta version active and when can we expect full fledged launch?

25 locations. Yeah, it is in the beta testing stage and it is performing as per our expectations. So right now full scale up will happen in about 45 days. We expect all our products to live on Project Cyclops by the festive period by maybe late November on all our products.

Considering that rural demand is picking up, to what levels can you see the retail book grow by the end of June 2024?

See our guidance, as per Lakshaya guidance, is that we will provide 25 percent CAGR. So that is what we are targeting.

Rural demand again is dependent on rainfall. Last year, rainfall was patchy in many parts of country, which hit the tractor business. This year though the predictions of rainfall has been good and the initial parts of the monsoon coverage, especially in the southern parts of the country has been okay, and there is some delay in monsoon activity in the western parts of the country as well as in the Northern parts of the country.

We are very hopeful that rainfall distribution picks up over the next few weeks, because that augurs well for rural demand this year and if rural demand is robust, I think the 25 percent growth rate will be quite achievable.

You are saying that retail book will grow by 25 percent CAGR - is it because you have most employees in rural area?

I think there is a lot of misconception that most of our employees are in the rural business, especially in the microfinance business. About 40 percent of our employees are in the microfinance business but we are building up our urban finance business also equally strongly.

Also read: MC interview | L&T Finance in tie-up with four fintechs for personal loans, vehicle loans, says Sudipta Roy

Which new markets are you planning to expand in?

We are expanding into a couple of new markets. We are starting a microfinance business. We started in Telangana and deepened our penetration in Western UP on the rural microfinance business front, especially through the rural finance business.

We are moving towards more penetration inside geographies where we had patchy sort of penetration and our micro loan against property business is also scaling up very well after the good response we received during our pilot in Tamil Nadu. Now we are scaling up in Karnataka, Maharashtra and Gujarat.

You have recently raised funds from JICA $125 million to finance women borrowers, how much out of this has been disbursed so far?

So we have drawn $50 million of the total $125 million we raised from the Japan International Cooperative Agency.

Do you have any fund raising plans through sustainability linked loans and social loan plans?

We expect to raise around $250 million in the current financial year.