Voice-based Deep Tech to Create Massive Differentiation in the Fintech Segment: Madhusudan Ekambaram, CEO, KreditBee

Deep technologies are becoming increasingly pervasive in the BFSI sector with the accumulation of data says Madhusudan Ekambaram, CEO, KreditBee

Supriya Rai
New Update
Madhusudan Ekambaram

With the COVID-19 pandemic accelerating the adoption of digital technologies, a humongous amount of data is being collected. This in-turn is giving Deep Tech the much-needed push. Madhusudan Ekambaram, CEO, KreditBee talks about the significance of Deep Tech when it comes to Fintech.


DQ: Deep Tech is a set of relatively new technologies. How has the COVID-19 pandemic impacted this technological shift in India?

Madhusudan Ekambaram: Prior to the pandemic, Deep Tech was already gaining momentum and seen as a significant focus area in the technology landscape. However, of all things that the pandemic has caused, it has bolstered the technological transformation, and the significance of Deep Tech in the purview has only increased. The application of AI/ ML has increasingly found utility in facets like facial recognition, face liveliness, optical character recognition and video KYC.


DQ: What are some of the trends we see in BFSI/Fintech as far as Deep Tech is concerned?

Madhusudan Ekambaram: Every facet of BFSI, especially the lending segment, from customer onboarding to the collection is witnessing significant change. The digitization of Customer Onboarding is complemented by the usage of Deep Tech in overall processes. The implementation of nuanced account aggregator infrastructure will only add on to the streamlining of processes and reduction in the turnaround time substantially. The underwriting process has seen major evolution with multiple data points being considered, enabling the efficient study of the cash flow to make effective credit assessment and proper decisions. The integration of ML to assess the potential risks is also boosting effectiveness. AI and ML are also greatly influencing the collection process, on both fronts - disbursements and repayments. This entire process is resulting in onboarding the right set of customers and building a stronger book with limited non-performing assets (NPAs).

Overall the major impact will be seen in the financial literacy and the resultant financial inclusiveness of the customers. The transparency of the lenders and favoring regulations will deem a healthy financial environment.


DQ: Similarly, what are the challenges companies face in implementing Deep Tech solutions?

Madhusudan Ekambaram: For any technology to work, the pre-requisites need to be appropriately aligned. In case of Deep Tech, it is the data which acts as the enabler of the technology. The effectiveness of Deep Tech depends heavily upon the quantum and quality of the data fed. To ensure good quality, requires heavy investments, and for companies like us, who have done it, the ante bears sweet fruits. This is particularly evident from our NPA scales and other performance metrics. Another challenge that a company might face, is the rapidly changing user behavior and external factors. This, however, only calls for enhancing the self-sufficiency of the technology model to be able to consider these dynamic variables.

DQ: Kindly shed some light on the kind of Deep Tech that is being used in KreditBee?


Madhusudan Ekambaram: At KreditBee, we have optimally operationalized the usage of AI/ML, to create a seamless customer journey. The deep integration of this follows in facets such as Face-match, Face-liveness, Digital E-Sign, Video KYC, Optical Character Recognition and working on SMS Variables. For the same, we use a mixture of in-house expertise and solutions provided by vendors like, Hyper verge amongst others.

In a business like ours, a robust underwriting system is highly essential. This is one particular aspect, where Deep Tech has an important role to play. Our scorecards used for customer assessment consider various data points such as personal profiles, social media thumbprints and spending habits. Deep Tech finds application here as the decisions made are not only based on vanilla regression but also on ML neural net models. Other Deep Tech applications such as decision tree to calibrate strategy, robotics and automation in the collection processes, vastly work towards improving our customer reach and creating a seamless customer experience.

DQ: What according to you is the road ahead for Deep Tech?

Madhusudan Ekambaram: While we do believe in efficient processes, in case of our technology adoption, we move slightly beyond the Pareto principle that 80 percent of effects are the result of 20 percent causes. Till now the Deep Tech integration has been revolutionary for us, but we believe that we are only scratching the surface, as we explore further opportunities towards the improvement of our operations and customer experience. The anticipated movements with respect to Deep tech application in the utility of data formats of image/text/video, will create massive differentiation in the Fintech segment. On the basis of this belief, we have started investing in voice-based Deep Tech, and this will find application in systems like AI contact/help center. The next few years are going to witness tremendous advancements in this respect and we are excited to adopt them.