India's economy is at an inflection point, manifesting unmistakable signs of recovery. However, the incumbent risk assessment practices come short in sustaining a resurgence in credit access, perhaps limiting the nation's aspirations to be on a high economic growth trajectory. Dated evaluation tools struggle to accommodate the explosion of data that characterizes modern businesses and individual lifestyles, failing to paint an applicant's risk profile in totality.
Consequently, as lenders are caught in a tricky balancing act of building a compliant, inclusive, and profitable loan book, thin-file customers lacking a detailed credit history are often denied funding, regardless of their alternate qualifications. Undeniably the prevailing situations in underwriting are a sharp contrast to other domains within banking and finance, where the proliferation of machine intelligence is consistently simplifying service access and adoption.
The reticence at a typical credit office is legendary. For decades the outcomes of its credit risk assessment and approval workflows have been influenced by static business rules that do not match the speed of business, long disbursal horizons, inability to weave information into a single source of truth, no integration with alternate databases, inadequate transparency, and high operational costs. Such systemic challenges apart, the lack of explainability of composite credit score developed through statistical regression using only up to 20 variables provides enough space for subjectivity and bias to creep into underwriting decisions. They vehemently deprive the underwriters of a broader view of the borrower’s profile.
Today, every BFSI institution is perching on massive databases on the existing to the bank (ETB) customers and can access even greater information sources available on-demand within and beyond the financial ecosystem. Further, advancements in collaborative AI and ML constructs for the banking and financial services that can intuitively operate on such data radically improve the quality and time to insights. In India, where at least 40% of the customers are new to credit, there is no reason why any archaic limitations should subdue the confidence of the lenders and continue to make them vulnerable to frauds and missed business opportunities undercutting credit growth in the country. The need of the hour is AI-infused assistance in credit underwriting that can assimilate disparate information streams into verifiable insights, supporting lending decisions.
Demographically, India is one of the youngest nations globally, with millions joining its economy each year. As a market leader in real-time analytics and decisioning, featuring more than 20 successful products in its portfolio, Perfios has been early in perceiving a steady shift in the average borrower persona towards new earners. The Perfios AI platform is the culmination of the company's efforts to decouple credit decisions from rigid policy-driven/rule-based underwriting methodologies and allow FIs to safely explore the business potential among the first-time borrowers or those with suboptimal credit scores. It is a SaaS platform that essentially seeks to redefine the relationship between machine intelligence and underwriting expertise in leveraging the ever-widening digital footprint of individuals and institutions as a reliable source of credit risk insights.
The Perfios AI platform adopts a holistic approach to credit scoring. As a decisive break from the legacy scorecard techniques, its AI & ML driven scoring engine pulls in data residing internally, not only in the bank's Loan Origination System (LOS), Core Banking System (CBS), and Trade Financing System (TFS) but also uses Perfios's industry-wide integration to access several external databases. It covers traditional and non-traditional sources, including demographic data, non-credit transactional data, financial data, credit/savings history, transactional data, economic data, environmental data, and data of alternate digital origins, making multidimensional credit profiling feasible.
For instance, a first-time borrower may not have an impressive credit score, an outright rejection as per conventional methods. Nevertheless, by using the same applicant's e-commerce habits, investment patterns, EMI contributions, or the utility bill payment frequency, Perfios AI can help the underwriter evaluate the case at a greater depth, discovering possible creditworthiness.
Computing at such a massive scale is made possible through the Perfios AI platform's neural network algorithms. It uses a common data model to compare the variables individually and reveal the underlying relationships between them. In contrast to statistical regression, the platform's neural network algorithms can accommodate an unlimited number of variables for arriving at the final credit score of the applicant in real-time and at a dramatically reduced cost. Additionally, it is a complete self-learning system, using ML algorithms to train and evolve continuously.
However, unlike most AI-driven credit scoring products in the market, the Perfios AI platform is not a Black Box. To guarantee a model level and loan-level transparency, besides the borrower score, it also provides the underwriter with detailed information on reasons for loan eligibility, the confidence level of the decision and recommends the approved amount, interest rate, loan term, and repayment frequency. With integrated reporting and analytics delivering end-to-end visibility into the outcomes, it promotes responsible and fair lending, eliminating subjectivity and bias from the decision loop.
Alongside the unmatched speed, depth, and accuracy that Perfios AI brings to the table, its exceptional scalability sets the platform apart from the competition. For too long, the underwriting process has been hardwired to the business rules and policies that are difficult to change and challenging to implement into the credit evaluation and approval cycles, leading to loss of opportunity in capturing the market.
However, Perfios AI is a modular platform that allows financial institutions to build and customize their business goals, rules, and policies on top of it, as per their specific products and service offerings. With this business rules around credit limit, fraud prevention, interest rates, and the bank's long and short term objectives can be approved by the corporate on the fly and configured dynamically into the platform with immediate effect, bypassing bureaucratic and technical complexities. It streamlines the functions for the underwriters operating downstream, minimizing the chances of goal divergence and policy breaches. Further, as a SaaS platform, Perfiso AI can be deployed in straight-through processing (STP) mode, completely automating repetitive functions within the credit approval cycle, improving turnaround time, and optimizing human resource usage.
At a time when banks are hard-pressed to expand credit coverage by treading through market uncertainties and volatile regulatory environments, Perfios AI might be the respite that underwriters have been looking for. While it brings an intelligent hands-on deck to work alongside human operators and prevent things from going south, it also enables the desired outcomes with greater transparency. The future indeed looks more collaborative for human expertise and machine intelligence with an expanded role of both in modernizing lending operations and transforming the credit exposure of the financial institutions.