Shilpa Bhandari is Senior Vice President and Global Head of Banking, Financial Services and Insurance at Birlasoft. Her goal is to have relentless focus on the clients by building strong teams, creating differentiated client value resulting in business growth, customer satisfaction & high team motivation. She has over 20 years of experience in the field and has previously worked with NTT Data services, Dell and Wipro Technologies.
Where do paper-based manual loan products with a 2-3 weeks processing time fit in today’s increasingly digitizing economy? The short answer: nowhere. Leading personal and commercial credit companies are redesigning their products from KYC to risk-management, loan decisioning, and underwriting with artificial intelligence and automation. Today’s customers – whether they are looking for personal loans or commercial credit – find a week-long loan-approval time obsolete. Moreover, leaders are upping customer expectations with delightful products that are personalized and reward them for making better choices. To survive in this cutthroat industry dotted with billions of dollars of untapped opportunities, it is time for lenders to adopt digital.
The State of Lending Business
In 2019, a subprime auto-lender was grappling with record low volumes of underwriting, and concerned senior management was busy formulating strategies to boost the volume and move capital faster. Several companies dealt with the same problem during the 2020 pandemic, but with added operational constraints. Manual home loan processing typically takes about 52 days, for example – and this year, loan approval rates have dropped by 50% on an average. In today’s digital economy, leading finance companies are capitalizing on vast amounts of consumer data and artificial intelligence to turn the tables around. Artificial intelligence in loan processing turns it instantaneous, cost-effective, convenient, and boost loan volumes in no time. In other words, artificial intelligence holds the keys to thriving in today’s cut-throat marketplace that’s getting fiercer with the arrival of alternative lending platforms and fintechs.
The Rise of Digital-led Automated Loan Processing
Traditional loan processing usually constitutes manual reviewing of documents for risk profiling, credit review, underwriting, servicing, and delinquency management. The manual nature of these operations results in significantly high costs for the lenders. It could manifest in the form of expensive employee-hours, lost time, and bad customer experience. However, in a digitally interconnected world, these processes have become obsolete and are a significant roadblock for the lenders who want to achieve high profitability in the lowest possible time. Personal and business credit customers alike want a seamless and lightning-fast loan approval and fund disbursement process. Observing this puzzle in the context of increasingly massive amounts of customer data, Fintech companies took the lead and gave birth to artificial intelligence and automation-based lending.
Driving Digital Transformation in Lending with Artificial Intelligence and Automation
Automated KYC Verification
Automated KYC processing leverages computer vision technologies. These advanced technologies use deep learning models for processing image and biometric data, and 3rd party cloud services for cross-validating data from multiple sources. Coupled with end-to-end automation of the KYC process, lenders can now reduce the traditional 2-3 days long, manual and physical process to an instantaneous, low-touch digital process. In addition to surging onboarding rates, automated, artificial intelligence-assisted KYC verification can eliminate human error and massively reduce onboarding costs.
Artificial Intelligence for Customer Data Remediation
As the variety, velocity, and volume of data accumulating at lending enterprises increases, industry leaders are already looking at artificial intelligence-based models and techniques to remediate customer data according to compliance requirements. Between 2008 and 2016, banks paid USD 0.32 Tn dollars in the form of compliance-related fines. Wherever paper-based, manual data remediation processes are a money and workforce effort blackhole, AI can help enterprises attain granular visibility and pinpoint missing values, contradictions, and SLA violations. Moreover, AI-based data remediation also paves the way for zero-error classification of accounts in line with specific compliance requirements by leveraging a centralized repository of enterprise data. It also helps the lenders improve loan application fraud detection and drastically reduce first party frauds.
Advanced Analytics for Customer Insights
While customers today show readiness to share their data for exceptional services, lenders must use more than a few factors in predicting customers’ eligibility for loans. Advanced analytics techniques such as multiple regression, high-dimensional classification algorithms, and predictive modeling can help lenders effectively predict the default probability of a case using hundreds of data points. Making such predictive models sharper requires highly optimized, self-learning machine learning models. The upside? By deploying such features, lenders can now offer pre-approved loans to customers, boost sales volumes, and further reduce loan processing times, while improving credit decisioning from a risk perspective.
Machine Learning for Credit Risk Analysis
As the diversity and volume of applicant profiles increases, traditional risk management techniques are falling short. To push a higher volume of sales and, consequently, revenue through the enterprise’s pipelines, cutting-edge machine learning techniques such as Artificial Neural Networks and Random Forests can help crunch high-dimensional and varied data within seconds to assess the credit-worthiness of prospective customers and fast-track the approval process. In addition to eliminating bias, usage of machine learning credit scoring models and automated loan underwriting is also empowering lenders to open up their products for more customers and fast – some lenders already see a significant improvement in default rates and consequent losses against their products.
Automated Loan Underwriting
An automated underwriting system requires lenders to effectively put the pieces together – that is, a robust and lean KYC process, high level of automation, and automated loan decisioning. Credit decisioning systems are already being reshaped by models that account for the minutiae of a customer’s data – like bill payment timings and spending habits in addition to traditional factors. Such models help enterprises build a low-touch underwriting process that brings human attention to outliers, thereby increasing the speed of application to disbursement while helping boost sales and save costs simultaneously.
Banks can augment rule-based underwriting with machine learning models to improve their accuracy. These machine learning models continuously train themselves as they evaluate customer data and creditworthiness. Through supervised learning, the bank can refine these models and better control the credit decisioning process. By applying automation to this decisioning process, underwriters can free themselves from cumbersome sub-processes like evaluating credit risk scores, reviewing all applications with the same rigor, and not utilizing the past data for quicker decision making. The underwriters can tweak the rules and models to ensure a balanced mix of man and machine intervention.
Final Words
The use of artificial intelligence in loan processing is catching up fast with banks and financial institutions. Early adopters of digital lending and lending automation processing systems witness 80% faster disbursement to customers, 30-50% reduced time-spend on decisioning, better risk profiling, and greater profitability, states McKinsey. In an increasingly digital world, paper-based, traditional lending processes are already becoming obsolete. High-touch manual processes are becoming unviable for many as a result of reduced workforce availability and high competitiveness. This shift presents an opportunity to reimagine the lending process – and push the lever to drive digital-led lending growth. Leaders are already looking at building embedded and integrated credit products and a trillion-dollar micro-credit opportunity. Those that are lagging must leap today to stay in the game tomorrow.