“AI-Powered Analytics Helps Predict Shifting Customer Behavior and Evolving Needs” Says Piyush Gupta

Innovation is rather a risky enterprise when it comes to Fin-Tech, all boiling down to how much market information one has. In the rapidly evolving and expanding industry it is highly imperative that one adapts fast, which in itself is a risk. Consequently, anticipating a product without giving up on the business needs of the fin-tech company is difficult if not impossible. However, PayU Finance thinks otherwise. Leveraging big data analysis and advanced AI and ML algorithms could reduce credit losses and drastically reduce NPAs. Analytics Insight has engaged in an exclusive interview with Piyush Gupta, Chief Data Scientist, at PayU Finance.

1.Kindly brief us about PayU Finance, its specialization, and the services that your company offers. 

We are one of the leading digital lenders in India and the NBFC unit of PayU. Our SBU, LazyPay, is India's preferred Buy Now Pay Later (BNPL) solution. We offer consumers a comprehensive credit suite, from deferred payments and BNPL at checkout to digital personal loans. We have since evolved to become India's credit super-app, offering online and offline payment solutions for low-ticket-sized purchases and EMI options for large-ticket-sized purchases.

We are looking to grow into a full-stack digital financial services platform to serve all financial needs of consumers through technology.

2. Kindly share your point of view on the current scenario of Big Data Analytics and its future. 

The widespread penetration and adoption of mobile internet have led to an explosion of data about consumers and their consumption behaviors and patterns. Mapping and understanding this data has allowed us to create solutions that are closely aligned with consumer needs, serving them in a much more personalized manner. Interestingly, because the usage of big data analytics is not just limited to big companies, every organization can employ the power of data analytics to generate better insights and build innovative solutions.

We are currently seeing a seismic shift in how big data is used. Going beyond analyzing data using batch processes, we are now moving towards real-time usage in decision-making. At PayU Finance, big data analysis plays a key role in building intuitive financial products that are one step ahead in anticipating consumers' needs. Our objective is very clear – we use AI and ML models in a fully integrated manner to offer a seamless experience to customers.

3.How is PayU Finance utilizing advanced analytics and big data? 

Our approach starts by defining the business problem and potential solutions in the appropriate context. We then utilize our in-house experimentation platform to conduct multiple quick experiments to understand the impact of the data and the value we can generate from that data.

We use AI and advanced analytics in several ways to help us improve our business strategies, processes, and operations and provide our consumers with a superior experience. For starters, we use advanced AI and ML algorithms to enhance risk assessment and underwriting to reduce credit losses without sacrificing business. Our core underwriting pillar, prudent lending, ensures that the debt obligation is within consumers' capacity to repay, and they are not overleveraged. Apart from bureau data, we also rely on alternative data to take a 360-degree approach to a profile, such as mapping location data and habits before offering a credit line. AI-powered analytics helps predict shifting customer behavior and evolving needs, to come up with relevant offers. Deep data wins the NPA game for us, while also transforming collections operations and improving performance at a lower cost.

Tech-enabled risk management capabilities help us in bolstering overall corporate governance and compliance. Summing it up, integrating analytics and data improves productivity, optimizes business operations, and helps with smart customer acquisition and retention through a superior and personalized experience.

4. What does your technology and business roadmap look like for the rest of the year?

Our data strategy is aligned with the lending business' ambitious goal of building a US$1.5bn loan book and a profitable combined credit entity by 2026. Our business roadmap is tightly linked to all our products and keeps the customer at the center. Going beyond underwriting and operations, we use AI and ML at every stage of our product journey. Our focus is to offer the right product to the right customer at the right time. This will enable not only effective and personalized marketing, smooth onboarding, and product usage but also a bespoke digital experience that is unique to each customer.

To further elevate the customer experience and improve credit risk, we are using state-of-the-art ML algorithms, graph neural networks, or adversarial networks. 

5.What are the concerns that organizations have before using Analytics? 

A company typically deals with many challenges when using analytics. One of the biggest, and most obvious challenges, is the quality of the data itself. Techniques by themselves cannot provide an edge to a company if the underlying data is "faulty". Another challenge that companies face is in selecting the right technology to store, process, and analyze data. This demands a significant amount of investment, often before you can generate any insights. There is also a constant need to create a delicate balance between data availability and data security. However, the biggest hurdle to successfully using analytics often lies in the organizational structure. Unless data is the central tenet, it is difficult to implement insights. A lack of support from upper management and employees could easily bring down the benefits of effective analytics.

These challenges can be even more pronounced for financial institutions where data is the starting point even before a customer can avail of your product, especially credit products. As advanced analytics combined with AI and ML techniques becomes a common practice among all players, one additional challenge that is creeping in is the ability to ensure that the models and insights aren't discriminatory in nature. To solve this problem, we, at PayU Finance, are designing, developing, and deploying responsible AI and ML models by not using any biased data or variables.  This ensures that every Indian is served with affordable and responsible financial products in the most frictionless way possible, irrespective of gender, caste, geographical location, age, other demographic factors, and even credit history.

6.What is the edge PayU Finance has over other players in the industry?  

At PayU Finance, data is the starting point to even determine the kind of product that we want to build or the experience that we want our customers to have. To enable this, we have used predictive analytics to create the biggest "completely pre-approved" base of 62 million consumers. The majority of these consumers are not served by any other financial institutions right now and that leaves us as the only player who can offer the financial products that they need. Access to proprietary data and the capabilities to make use of this proprietary data is what sets us apart from competitors. We also have the cost of fund advantage as our NBFC unit is very well-rated in the market and bolstered with significant equity.