Qorus Banking Innovation Awards 2023

Preventing Non-Performing Assets (NPA) by Proactively Forecasting Accounts High Likely to Default

22/09/2023 Banking Innovation


Our AI-driven system transforms risk assessment, automates NPA date prediction and amount required for NPA Resolutions, and scales across business domains, setting a new industry standard.

Innovation presentation

The bank has a large number of working capital customers, and its customer base is rapidly growing. Regulations require the bank to closely monitor the activity of these accounts and flag off accounts that are at high risk of becoming NPA. The Bank proactively engages with these high-risk customers to regularize their accounts before they are classified as NPAs.

The earlier process lacked predictive capability to determine the likely NPA date and categorize customers by risk level based on their profiles and interactions with the bank and had high mis-classification rate. This limitation hinders the organization’s ability to effectively prioritize collection efforts.

Moreover, the current process of identifying NPAs is entirely manual and heavily reliant on Excel, taking almost 4-5 hours daily to identify potential cases.

To overcome these challenges, we created an AI-driven analytics platform supported by a statistical machine learning model. This system accurately identifies NPA cases, leading to reduced NPAs and fewer customer complaints. It's the first fully digital AI platform of its kind, enabling timely NPA prediction and proactive measures to prevent them.

Loss Prevention:

This solution helped the Bank save INR 270 crores in NPA losses in July 2023 by proactively predicting high-risk accounts, enabling the bank to resolve 92% of accounts in advance (up from the previous rate of 55%).

Enhancing Customer Experience:

Previously, the Bank contacted customers daily for payment updates, leading to an increase in complaints about frequent calls. In certain instances, HNI clients were erroneously flagged as NPAs, resulting in complaints to the RBI.

After implementing the platform, complaints decreased by 65%, and customer calls reduced by 38%, saving resources and boosting productivity.

Scalability to Other Verticals:

In August, we expanded the application to various verticals (Agri Loan and Loan against Property) and collectively saved INR 689 crores from going into NPA. On some days, we achieved a 100% resolution rate, preventing all accounts from becoming NPAs.

Uniqueness of the project

1. Automation of Risk Assessment: The platform has automated the manual NPA assessment process, significantly reducing the turnaround time from 4 hours to just 15 minutes. This newfound efficiency has allowed the bank to focus more on critical cases and take proactive measures to prevent potential NPAs.

2. Predictive Analytics for NPA: By leveraging ensemble machine learning techniques such as Random Forest and XGBoost, the platform accurately forecasts which accounts are most likely to become NPAs. Moreover, it provides insights into when this transition is expected to occur. This foresight helps the Bank proactively approach these critical cases, minimizing the impact of NPAs and reducing customer complaints.

3. Customer Profiling & Segmentation: Based on the customer's overall relationship, product holdings, as well as payment performance with the bank and credit bureau, the platform segments customers into High, Medium, and Low-risk categories. This segmentation enables Bank to allocate resources more effectively, directing efforts towards high-risk cases while efficiently handling low-risk cases through digital channels.

4. Prescriptive Layer: The platform is designed to prescribe measures to prevent NPA and prescribe the amount required to regularize or prevent an account from going into NPA. This enhances the bank's risk management strategies, ensuring that the right actions are taken to mitigate NPA risks effectively.

5. Scalability: The platform is not limited to a specific product or business vertical. It has already started scaling across multiple business teams within the bank, including Loan against Securities, Gold Loan, Agri Loan, working capital Loans, to prevent Out of Order NPAs across various domains. This scalability ensures that the value of our innovation can be harnessed across the entire organization.

In summary, this AI-driven utility is a ground-breaking innovation for the Bank, aligning with regulatory requirements and significantly reducing NPAs and customer complaints. Its ability to automate, predict, profile, and prescribe actions not only saves time and resources but also enhances the overall effectiveness of risk management in banking, setting a new standard for AI/ML-driven financial risk assessment.

Want to keep reading?

Become a Qorus member to get access to all our innovations

Interested in learning more? Speak to Boris, Qorus's Content Lead

Qorus has a library of almost 8,000 innovation case studies across critical areas like customer experience, sustainability, marketing & distribution and more that can be used to inform your decision-making.
Contact us

Related innovations

22/02/2024 Insurance Innovation

Master Actuary: AI For Actuaries

A new green pricing era ? : Master Actuary is so powerful that it can assess driver risk based on...

21/02/2024 Insurance Innovation


SightBit is an Israeli company specializing in autonomous artificial intelligence systems for detecting and predicting hazards in open waters. SightBit’s...

15/02/2024 Insurance Innovation

An expert weather model for a fast and efficient claims management

To simplify the management of weather-related claims, Société Générale Assurances has integrated a solution that enables its claims handlers to...

15/02/2024 Insurance Innovation

Cost-Effective Coverage with High Customer Satisfaction

At US Automotive Protection Services, we stand out by offering cost-effective car coverage solutions that don't compromise on quality. Our...

Related news & insights

Digital Reinvention
29/02/2024 Article

From black box to open playbook: Bringing transparency to AI in banking

Make the transition from opaque, closed-off strategies (black boxes) to clear, understandable tactics (open playbooks) through explainable artificial intelligence (XAI).

Digital Reinvention
28/02/2024 News

Klarna's AI assistant transforms customer service landscape

Klarna's recent unveiling of its AI assistant, powered by OpenAI, has already made significant waves in the customer service realm.

Open Finance
27/02/2024 News

Catapult: FundTech 360° by The LHoFT Foundation to boost fintech innovation in Luxembourg's fund sector

This initiative aims to bolster Luxembourg's standing as a global fintech leader, focusing on enhancing the fund sector's governance, cybersecurity,...

Digital Reinvention
27/02/2024 News

Cash on call: CaixaBank and Correos partner for convenient delivery service

Exclusively available to CaixaBank clients, this innovative offering allows individuals to request cash deliveries ranging from 60 to 500 euros...