How PayU Money using data science technology?

Posted by Amit on January 14th, 2021

In the technology world right now, data science is a buzzword and, for good reason, it represents a big step forward in how machines can learn. We are the best institute for data science in Delhi ncr because our vision of offering courses concerns various types of applicants with different working experiences. 

In order to understand their customers and make decisions on the basis of insights, companies in the financial credit field are becoming overzealous about data science and AI models. Data science has now become the backbone of any BFSI player and for companies such as PayU and its line of business, in particular, it plays an important role.

What is PayU Money?

PayU Money is the Indian fintech business that supplies online merchants with payment technology. It enables online companies, via payment methods that can be incorporated with web and mobile apps, to accept and process payments. The PayU Group has operations in 16 countries in Asia, Central and Eastern Europe, the Middle East, Latin America, and Africa.

The Indian operation - PayU India - is one of the country's top three payment gateway providers with a market share of more than 30 percent, comprising more than 300,000 dealers. PayU India offers more than 70 methods of online payment and aims to balance the needs of retailers with the way customers shop and pay.

How PayU uses data science technology?

Global online payment services company PayU announced in August 2016 that Citrus Pay, the leading Indian payment technology firm, will become part of its Indian business. The transformed PayU India activity had to provide secure, sensitive payment services in order to survive in a competitive market. With the merger of PayU India and Citrus Pay, the company needed to integrate data from both parties to optimize the use of data to make data-driven choices.

The data science team of PayU India estimated that the company would need to quadruple its current infrastructure while consolidating all data sources into a single database. The emergence of new and enhanced AI and ML technology has the potential to allow seamless access to credit, digitally link businesses with customers and on-board new customers.

As a testament to the importance of data science, AI and ML are used at PayU during the life cycle of a client. As well as possible fraud, PayU uses it to control credit risk. A few examples of use-cases where they use ML/AI to improve client experience are frictionless onboarding, service automation, and avoiding excessive selection efforts. Using data science and a data platform, PayU responsibly creates customer insights that help our merchants and the ecosystem grow their businesses.

A collection of models are used together at PayU to assess a customer's 'worthiness'-both their capacity to pay and their ability to pay. Similarly, the company has ML models to predict customer churn, the probability of auto-payment by a customer who skipped payment, the likelihood of a customer completing KYC, etc.

Their customer support chatbot uses a home-grown NLP model generated using previous data from human service. PayU not only uses conventional data sources such as repayment behavior of customers, banking, and office information, but also how a customer communicates with the site, applications, spending habits of the customer in their payment services, etc.

Using data to generate useful insights –

AI/ML models flourish only when the data is collected and stored by a stable, reliable, and scalable framework. PayU has a dedicated team to design, develop, and manage data systems based on the cloud to run and train different models. The required framework performs transformations in real-time and batch data intake and makes it available both in a data lake and in a centralized warehouse.

Automated functions of the platform include running ingestion pipelines, data cataloging, quality control, etc. The company's data scientists, market analysts as well as internal and external reporting and business intelligence use the data lake and data warehouse. 

Understanding customers’ credit line –

PayU also takes a 'low and expand' approach, beginning its relationship with the customer through LazyPay, our Buy Now Pay Later solution, through a deferred payment experience of just a few hundred rupees. They keep growing the sum of the loan by monitoring data and consumer insights, as they see timely repayments from these customers. They aim to check earnings, commitments, etc. by using more conventional banking means to offset risk as the loan amounts get larger. 

Then, without the need for unnecessary documentation and offline procedures, they digitally represent a huge credit-worthy population in India. They're supported by data science. Scale the direction of a customer from deferred payments to a big ticket personal loan. To incorporate this paradigm, PayU uses cloud-based tools, which enable the team to monitor and compare the performance of multiple iterations of a model. 

In western markets, not allowing AI-induced bias in credit models is already a major factor and will become so in India as well as with changing regulations. A responsible AI structure for their credit models was developed by PayU. First, they ensure that in deciding the credit line, the models do not use any data or variables that can slip prejudices of gender, religion, political ideology, etc. Secondly, the ML algorithms that are implemented are interpretable, so why anyone was denied credit can be clarified. Finally, before deploying any new model to assess if it is stable and bias-free, they have human interference for credit models. 

Conclusion-

PayU will bring synergies in data and talent across their payments and credit businesses. They can look forward to using the Payments data to determine the affluence of a customer, which can then be used to pre-approve a large segment of customers for a certain credit limit. In addition, to cross leveraging their data and team, they are also exploring newer AI techniques such as knowledge graphs and embeddings to better predict credit risk.

Like it? Share it!


Amit

About the Author

Amit
Joined: April 6th, 2020
Articles Posted: 4

More by this author