NBFC leverages artificial intelligence and machine learning to automate business processes
Over the past two years Hinduja Leyland Finance has focused on activities beyond our traditional new vehicle finance business and we are now trying to tap into the other life cycle of a vehicle, namely the operating life cycle and the resale market of a vehicle. To learn more about the phygital model of NBFCs, Shruti Jain of Elets News Network (ENN), interacted with Kunal Kathpalrisk manager, Hinduja Leyland Finance.
Q. How does AI and ML contribute to the ongoing evaluation of the underwriting and risk model for NBFCs?
Rep. NBFCs in India are required to perform KYC and credit assessment as part of the credit underwriting process. Traditionally, these were done based on documents provided by the applicants and then a credit underwriter would manually interpret them based on the credit policy. This method presented challenges such as scalability, evaluation errors, cost and standardization, etc. This model also had its own share of limitations like input errors, outputs had more returns than decisions.
However, over time, NBFC learned this art of lending and turned it into a science, giving it the agility to adopt AI and ML for this customer segment.
With increasing digital footprint in the new era, NBFCs in India are leveraging Artificial Intelligence (AI) and Machine Learning (ML) solutions to automate
their key business processes such as customer onboarding, decision making, risk management, etc.
According to a recent study by FICCI and PWC, 83% of Indian financial organizations say that AI helps improve customer experience.
These models are prepared in such a way that they are improvised based on historical data and current trends. Credit model management teams continuously review model adjustments based on correlated macroeconomic indicators. The performance of these models is compared to the organization’s risk appetite. It is therefore a continuous evaluation process.
Q. As emerging technologies disrupt NBFCs, what risks does your organization face in onboarding traditional customers during the journey?
Rep. As India has embarked on the journey to increase its digital footprint and grow the banking population by expanding its banking network to tier 3, 4 cities and rural sector. There are still customers who are new to credit with a small digital footprint, hence the reliance on manual onboarding for these customers.
When onboarding these customers due to the limited digital footprint and lack of data in some areas, credit underwriters face challenges such as –
Subjective parameters – there is still a heavy reliance on documentation for some of the settings in the dashboards which are populated by personal judgment.
Human interpretation – lack of authentic data source that can help to assess viability assessment and borrower cash flow human intervention is required to interpret the data.
Scalability and decision time – Given the huge reliance on human judgment based decision-making process, this will always cause scalability issues and require time-consuming efforts as each decision-making parameter has to be validated manually.
However, as the digital footprint grows, traditional models are also evolving towards “phygital – combination of digital and physical” subscription models.
Q. How is RPA used in the risk management section? How does it prevent fraud?
Rep. Automation of robotic processes is an automated technology that can record actions performed by humans on a computer, and then this process can be done without humans.
RPA helps reduce overall risk in four ways:
1. Minimize human judgment – An RPA robot performs tasks without any human faults such as biases, variations, errors or fatigue.
2. Simplification of credit reporting – When applying RPA in the loan origination system, processes such as verification, conversion, and KYC validation become more streamlined, thereby streamlining the overall credit assessment of applicants.
3. Compliance Factor – An RPA will be different for different types of products.
This ensures that any product-specific changes are updated in the RPA system, reducing the time needed to recreate separate systems for separate products.
4. Standardize the process – An updated version of RPA can ensure that your business will be able to keep up to date with all the necessary requirements. The right RPA automation tool would be agile and mitigate risk by allowing systems to make room for any new changes to occur and deal with complexity.
RPA prevents fraud in the following ways:
1. Reassess current processes
Financial organizations can program RPA bots to examine current and historical financial transactions to find anomalies and typical patterns that may indicate illegal or fraudulent activity. RPA will require the financial institution to investigate, document, and evaluate current processes, leading to deeper insights and identifying high-risk areas.
2. Identification of vulnerabilities – RPA bots can be automated to review current transactions and compare them to historical transactions to assess a pattern in a timely manner, helping to identify uneven patterns that may expose anomalies that may be illegal or fraudulent activities .
3. Accelerating Fraud Investigations –RPA gathers data from multiple sources and then analyzes it faster, allowing investigators to spend more time solving fraud cases.
Q. How does Hinduja Leyland diversify risk across sectors?
Rep. Hinduja Leyland Finance has diversified its portfolio into various categories such as commercial vehicle finance, construction equipment finance, personal vehicle finance, loan against property and affordable housing finance.
All of these products have their own set of customers and markets with different behavioral patterns, allowing us to design our products around a different risk appetite for each.
What are your plans for 2023?
Rep. As a company, we are on a continuous journey to achieve our mission of “to be among the most preferred financial service providers in India for all our stakeholders (customers, partners, employees, shareholders) “
Over the last 2 years HLF has focused on activities beyond our traditional new vehicle finance business and we are now trying to tap into the other life cycle of a vehicle which is the life cycle operating and vehicle resale market. For this we have launched platforms like GRO digital in partnership with Ashok Leyland where GRO will manage (and in the process facilitate) various operational aspects of the vehicle operating life cycle (such as load matching, fleet, fastTag, fuel refill, etc.) and HLF will offer credit products around this, such as bill discounting, tire credit, fuel credit, etc.
Also, HLF through its subsidiary Gaadi Mandi has created an online platform for the sale/purchase of used vehicles. This has contributed to a better fulfillment of the resale of repossessed vehicles by HLF and other lenders (with whom we plan to link up) by removing middlemen.
These initiatives, however, are in their infancy and will help us grow faster by mitigating risk (helping the client grow revenue through the GRO platform and Ring Fence cash flow, as the cash flow and payments will go through these specific platforms/systems) and improved fulfillment and cost savings through Gaadi Mandi.
Therefore, we will continue our journey towards achieving our mission statement by digitizing more and more processes as our customer’s data footprint grows and create an ecosystem of ease of doing business by supporting the customer while throughout its life cycle.