mPokket is an Instant Loan App that provides loans to college students and young working professionals. mPokket started back in 2016 with the vision to improve the lives of those who need quick cash for an urgent need but do not have access to credit cards or other means of formal borrowing. The entire process is online and usually takes thirty mins from downloading the application to the disbursal of the loan amount as long as the information provided by the user is correct.
Andy Sen is the CTO of mPokket. In an interaction with The Tech Pod, Andy speaks about the role of AI in fintech industry. Read more!
Tell us something about yourself and what does your company do?
mPokket is an Instant Loan App that provides loans to college students and young working professionals. Borrowers can avail loans ranging from Rs.500 to Rs.20,000 and the amount is instantly credited to the users’ bank account or digital wallet. One can choose to repay in up to 3 months along with a nominal interest charge. The borrowing limit can increase over time based on multiple factors, including timely loan repayments.
mPokket started back in 2016 with the vision to improve the lives of those who need quick cash for an urgent need but do not have access to credit cards or other means of formal borrowing. The entire process is online and usually takes thirty mins from downloading the application to the disbursal of the loan amount as long as the information provided by the user is correct.
Unlike other instant loan apps, we do not restrict our services to a limited set of cities or colleges. One can be from any college or be working in any city or village in India and we will provide you with a loan as long as you provide valid documentation. Our ease of use is what sets us apart and has helped us become the best instant loan app in India. Our goal is to increase financial inclusion by making financial services available to the underserved and new to credit population.
I am currently the Chief Technology Officer of mPokket. As the CTO, I am responsible for overseeing the end-to-end management of the technology function at mPokket, which involves product development and delivery, continuous product innovation and ensuring its high scalability. Furthermore, I am also working on building a robust technology team, from the ground up.
I joined mPokket in April 2019, with the aim to scale the platform to not just handle the current user-base of 10 million but also expand the base to 50 million users across the country. Having already begun incorporating AI and ML techniques in modules like signup, credit scoring, and fraud detection, I envisions levitating mPokket, within the next five years, to a stage where both regulatory and functional complexities of the business have been automated to a great extent.
Why AI is the Future of Finance?
Deterministic if-then-else models have difficulty scaling due to the increase in factors involved in taking credit related decisions. Once upon a time your take home income was the only factor, now we have so many of them – expenditures, type of housing & ownership, type of vehicle owned, nature of employment and so on. Trying to come up with all the combinations of factors and their respective ranges across the population segments is tedious and fragile.
It is preferable to actually use machine learning models along with real data (obtained by sampling a % of your user base) to achieve scale both in terms of factors and user population size.
What are the use cases of AI in mPokket? Describe how tech is playing role in each of the following functionalities
• Credit Decisions
As mentioned above, we use multiple variables in deciding the credit limit. As our users have no significant credit history, depending on external credit agencies (like CIBIL or Experian) is not as valuable as having a custom in-house built credit model. Considering the large number of factors involved in the credit decision, this is something we handle using ML & AI.
- Customer Engagement and Management
There are two parts to this – we use a lot of automation to deal with customer management, including speeding up our on- boarding and KYC processes. However, having said that, we still feel that our users actually prefer talking to our agents directly and hence for engagement we provide both email and telephone support.
Chatbots sound good in theory, but if you’ve ever used one during your time of need, you’ll realize how far they still have to go in order to be truly useful. We prefer giving the human touch and we have found that our users also overwhelmingly choose this option.
- Credit Scoring
Please see above note on Credit decisions. In addition, our user’s score change on a daily basis – based on events which may have transpired over the day.
There are multiple checkpoint we’ve integrated – all the way from detecting if a KYC document was digitally altered (“photoshopped”) to ensuring loans are taken in the correct bank account (we’ve tied up with our banking partners for this).
Is Technology enabling fintech players to actually reduce due diligence costs drastically? How?
What due diligence costs are going up overall, technology has helped fintech players like us scale well. We are able to serve 10-20x the number of users using the latest available tech than if we’d use older approaches. In addition to the scale, there’s also a significant improvement in “velocity” – where we’re able to quickly bring new features to market.