FlexiLoans adopts AI and deep learning tech in its lending business

As Indian financial services started accepting the concept of Artificial Intelligence (AI), startups such as FlexiLoans have started applying it to their daily lending business.

“In the past one year, the company has tried to create a machine that can lend,” said Abhishek Kothari, co-founder of FlexiLoans.  “If we add more humans in the lending chain, more frauds happen.”

The company usually give out 800 loans in a month, and in order to curb the possibility of negotiating loan terms and reduce the risk of misaligned incentives among other concerns, FlexiLoans has created an algorithm.

“Out of 800 loans in a month , around 300 of them are done via an algorithm automatically, No human touches it until it has been sanctioned,” said Kothari.

As per a NASSCOM-CMR survey (Artificial Intelligence for Banking, Financial Services and Insurance Sector, 2018), the adoption of AI in the financial services sector  would offer a better customer experience, the automation of back-end business processes, and effective compliance and risk management.

“We are trying to create algorithms which can actually behave like humans. Our chatbot appears as if you are chatting with a human, but it is a machine on the other side. It will keep asking you questions. We have also embedded an algorithm to understand documents. It has a vision of it. The algorithm took 6 to 9 months if hardcore AI training,” said Kothari.

FlexiLoans has set up a lab that experiments with deep learning and technology application. Called ‘Delta Lab’, it creates algorithms and vision techniques for future lending opportunities.

“No one in the fintech industry is actually thinking at the landscape of artificial intelligence. People are still doing automation and we moved further away from that,” Kothari said.

He said that although AI today is used for simple techniques, but tomorrow, the same technology can be used to detect fraud on a bank statement where the person has added an extra zero on it by identifying the additional pixels generated by the numeral.

While comparing humans with machines, he said, “Machines will always be objective, which means the upside is they will always be predictable. The downside is there are some people in the borderline who deserve credit, the machine will say no, you don't deserve one. At the end  of the day you are not trying to solve every man's problem, you are trying to solve a population problem.”

Kothari said that he cannot determine how much the company has saved by adopting AI as there is no return on investment (ROI) determined on it.

According to reports, the government of India has allocated Rs 3,073 crore for the Digital India programme which includes focus on research, training, skill development in robotics, artificial intelligence, big data intelligence and quantum communications.

Many investors are also showing interest in AI-related startups. According to NASSCOM, close to 400 AI startups in India secured an investment of $150 million in a span of 5 years. The overall funding for AI has increased to $73 million in 2017, up from $44 million in 2016.