Advanced Machine Learning Techniques to Do Real-Time Credit Assessment

ePayLater uses advanced machine learning techniques to do real-time credit assessment by leveraging data such as buying patterns with the

Ashok Pandey
New Update
Machine Learning, Credit Assessment, ePayLater, digital payment

The tech advancements have changed our lives in various ways, and the banking sector has no exception. Forget the queue for cash deposits and withdrawals, or applying for loan and credit card. This is the story of past, today, we have access to all these on our smartphone. And the next step is getting credit without a credit card.


ePayLater, a digital payment solution offered by Arthashastra Fintech Pvt. Ltd., enables a "Buy Now, Pay Later" product for frequent online purchasers with an interest-free credit term of 14 days. It is recognized as a startup by the 'Department of Industrial Policy and Promotion' of the Government of India's Ministry of Commerce.

We spoke to Akshat Saxena, Co-Founder, ePaylater, to understand the fintech industry and the technology propelling its growth.

The new opportunities for the Indian fintech industry in 2019


Fintech firms are breaking new ground in the formal finance sector through the innovative and dynamic use of technology in the lending process. Currently, the availability of credit at the point of sale is more or less absent for a vast majority in India, especially when the latter is digital e.g. e-Commerce portals. A number of people, despite being creditworthy haven’t had a way to instantaneously acquire a credit instrument at the point of sale itself and conveniently transact then and thereon. However, with a continuous ingress of a younger working population coupled with poor penetration of credit cards, the need for digital credit is on an unprecedented rise. With a credit card penetration of less than 3% in a population of over 130 crores, we see a huge headroom for growth.

We also see business credit as a major thrust area, as there lies a huge unmet demand for easy and formal credit. India has over 50 million SMEs and less than 10% have access to credit. Many of them are centered around rural and semi-urban regions and have nil to little history on credit bureaus. As a result, a vast segment of these is still unbanked and untouched by any form of formal/bank credit. Adding to their woes - many lack sufficient documentation which does not make the process any easier. We believe solving the said problem would unlock the true business potential of micro-entrepreneurs across the country. Having experienced the green-shoots of this welcome change ourselves, we feel propelled to drive this with added focus and intensity.

Another big game changers would be UPI and QR based transactions. Both have really made payments seamless and is being adopted by customers across the country. As we speak many fintech companies and banks are laying down the infrastructure to facilitate UPI and QR based transactions because it is convenient and cost-effective for both the merchants and customers


The challenges for the fintech industry in 2019

In the coming years, consumers for Indian Fintechs are likely to be relatively new to the internet, largely from rural India, and with much less familiarity with technology and formal financial channels than the typical urban user of today. This is why we need further improvement in the digital infrastructure of the country. Fintech companies in India must evaluate how they can make their products and services useful and accessible for this new demographic, as this is where their growth and success in the long term reside.

How is AI/ML helping to do credit assessment?


ePayLater leverages data science to utilize alternate data to assess the creditworthiness of an applicant. ePayLater uses advanced machine learning techniques to do real-time credit assessment by leveraging data such as buying patterns with the merchants, digital footprint, social media information, device information etc. We leverage sophisticated algorithms that have the capacity to cater to wide and differing audiences in an inclusive way. Thus, we are able to extend instant credit to even those individuals who have no history of availing credit before.

For instance, ePayLater works with different merchant partners that cater to very different target groups and markets. IRCTC sees traffic from the biggest metros and the smallest towns and gets customers of different ages that often lack credit histories. In spite of these challenges, ePayLater’s algorithms have been able to determine their creditworthiness using alternate methods to gather and analyze the data available.

Risk assessment to estimate the creditworthiness

While providing credit to customers ePayLater goes through a process of risk assessment to estimate the creditworthiness of a prospect. Traditional systems relied on historical data like credit history, bureau (e.g CIBIL) score and income to understand the risk associated. However, in a country like India where the credit penetration is almost negligible, much of the population has little or no credit history. Also, historical data is not always an accurate standard to predict future behaviour. ePayLater uses advanced machine learning techniques to do real-time credit assessment of its customers.

This real-time assessment by ePayLater ensures that a customer’s credit approval status is shared instantly. The approval rate is also higher than traditional banks due to leveraging of alternate data along with traditional. This means a customer can totally avoid the hassle of standing in long bank queues for just getting his/her credit card approved.

epaylater digital-payment