With HyperVerge OCR, LTFS was able to reduce operational costs by 50%

INTRODUCTION:

L&T Finance (LTF) is one of India’s largest financial companies. It has always strived to work towards financial inclusion and has consistently expanded its services and offerings to the underserved regions of the country. In other words, it caters to Bharat – where most of the Indian population lives. Over the years, LTF has empowered numerous individuals, farmers, and small businesses to reach their financial goals through its innovative financial products. The company aims to be a Fintech at scale by 2026, bridging India’s digital and financial divide.

THE PROBLEM:

LTF wanted a highly automated AI solution for validation.

In L&T Finance’s operational landscape, the journey of disbursing two-wheeler loans relied entirely on manual processes for extracting, classifying and verifying non-standard documents such as insurance papers, MMRs (margin money receipts), and invoices.

As a result, LTF faced several challenges: their manual system demanded representatives to manually input and upload details from invoices, MMRs, or insurance documents for each submission. After submission, the backend ops team had to manually review all uploaded documents, checking for errors and potential fraud. Example: Discrepancies in names or insurance details matching the customer’s information.

Given the thousands of applications LTF receives each month, this process proved excessively tedious and time-consuming, leading to increased operational costs, extra effort, and longer turnaround times.

Hence, LTF wanted to eliminate the manual process for validating these non standard documents. Validation is the process where a user’s identity is verified against the data extracted from a set of documents.

What LTF was looking for:

In essence, LTF sought an AI solution that achieved  the following objectives:

• An automated solution for extracting and categorizing non-standard documents.
• A streamlined process devoid of manual intervention at any stage.
• Enhanced user experience through quick turnaround times.
• Efficient backend operations with minimal manual oversight.

THE SOLUTION:

The LTF team identified the necessity for an automated solution over a human-led approach and chose to partner with HyperVerge for its validation process.

Following the implementation of HyperVerge OCR in the LTF two-wheeler journey, the turnaround time became almost instant, and HyperVerge OCR efficiently extracted all fields from non-standard documents.

HyperVerge OCR:

Our OCR engine effortlessly extracted all essential fields from insurance documents, MMRs, and invoices while cross-verifying them against the entered details, effectively identifying any errors.

Achieving over 90%+ accuracy, HyperVerge’s system was able to ensure a higher green channel percentage of 70%, i.e. these were approved in a straight-through-process (STP). The remaining 30% were flagged for manual review due to one or few fields.

HyperVerge OCR was also able to reduce the operational costs by offering nearly an instant TAT, Our AI model validated thousands of applications monthly, a task that would otherwise demand 2x the labor and expenses for manual validation.

IMPACT CREATED:

• 70% Straight through processing.
• Operational costs reduced by 50%.
• Accuracy rate: 90%+

WHAT’S NEXT?

LTF aims to streamline the validation process across farming loans, personal and consumer loans, and MSME loans by integrating HyperVerge OCR. This automation includes the handling of insurance documents, partnership authorization letters, electricity bills, and various non-standard documents. 

LTF is currently partnered with HyperVerge for customer onboarding. Continuously innovating in complex digital lending products while prioritizing customer needs remains central to LTF’s mission. Both organizations share a common goal of enhancing financial inclusion in India, paving the way for ongoing innovation and positive societal change within the financial services sector.