Fibe Accelerates Business Growth and Transforms Digital Lending on AWS

Non-banking financial companies (NBFCs) are integral components of India’s financial system. As of October 2022, over 9,500 NBFCs were registered with the Reserve Bank of India, with a cumulative asset size of INR 42.05 lakh crores (approximately $508 billion). Fibe, previously known as EarlySalary, is a financial technology company that offers instant personal loans and salary advances through its digital platform. Its innovative technology quickly assesses creditworthiness, notifying applicants within 1 minute about the status of their loan application. If approved, the loan amount is swiftly transferred to their bank account, providing convenient and accessible credit options for short-term financial needs.

Since its launch in 2015, Fibe has expanded its product line to include long-term loans with tenures of up to 36 months. Additionally, the company has ventured into buy now, pay later services for the education and healthcare sectors. Customers now have the option to pay for medical treatments or school fees in equated monthly installments (EMIs).

With more young professionals entering the job market and India’s middle-class population projected to reach 583 million by 2025, significant growth opportunities were evident. Recognizing this, Fibe sought to deeply integrate machine learning (ML) into its platform to offer faster and more convenient services to its customers. However, the company wanted to ensure that these advancements did not lead to increased administrative overhead. Anil Sinha, chief technology officer at Fibe, says, "We operate on a principle of balance—enhancing services while keeping costs in check.”

Fibe was born on AWS. Sinha says, “AWS has been incredibly supportive from the start, offering extensive training and providing assistance. Leveraging AWS Support saves us valuable time and ensures we can quickly resolve any issues without wasting effort on problem-solving.”

Since the launch of the Fibe app, Fibe’s IT team has been using AWS Lambda to streamline development and maintenance processes. AWS Lambda's serverless functions automatically handle code execution requests, simplifying processes like scaling instances to meet higher computational demands. “We realized that serverless technology was the best solution to minimize management overhead for our critical infrastructure,” says Sinha. 

To enhance the loan application process, Fibe incorporated ML into its app while maintaining a simplified management approach. To achieve this, the company implemented Amazon Textract, a service that utilizes optical character recognition (OCR) technology to extract text and data from Know Your Customer (KYC) documents. “Amazon Textract helped us automate things like face matching, without the need to handle infrastructure management,” says Sinha. Additionally, Fibe uses Amazon Comprehend, a natural language processing service, to extract relevant information from income statements and other application documents, reducing administrative complexity.

After implementing these solutions, Fibe turned its attention to enhancing platform security and detecting fraudulent activity during the application process. The company worked closely with its local AWS team and deployed Amazon GuardDuty, which utilizes anomaly detection to quickly identify threats. Additionally, Fibe started utilizing AWS Config to continuously evaluate the compliance of its AWS resources.

Next, Fibe began leveraging Amazon Neptune to detect and prevent fraudulent loan applications, allowing them to create fraud graphs without incurring infrastructure costs. The solution utilizes data from a data lake on Amazon Simple Storage Service (Amazon S3) and is based on relationships between different entities, including loan details, financial transactions, device fingerprint, IP address, and other relevant data points. This interconnected network helps Fibe to explore and identify fraudulent activities. Sinha affirms, "The deployment process was seamless, and we successfully built our data lake infrastructure on AWS within a span of just two weeks."