The problem EmpowerCredit solves

Our project addresses the financial challenges faced by 90% of India's population working in the unorganized sector who struggle to obtain loans due to poor or non-existent credit scores (CIBIL scores). These individuals often resort to high-interest loans from informal sources during emergencies.

To solve this, we will develop an application that uses an algorithm and machine learning (ML) model to analyze user’s bank messages and other financial parameters, generating a reliable credit score. This will enable access to loans at lower interest rates. Additionally, we will implement a calling feature to help users find suitable government loan schemes, addressing the issue of missed opportunities for financial assistance. Our business model includes a loan repayment portal, where a small service fee is deducted from the interest.

Challenges we ran into

Integrating with bank APIs posed significant challenges due to the varied and inconsistent formats of PDF bank statements, where each row contained disparate columns and types of information. Extracting transaction data into tabular format proved particularly daunting given the diversity in statement layouts across different banks. Additionally, existing OCR technologies utilized for reading these statements did not consistently and accurately generate transaction details promptly, leading to suboptimal latency.

To address these complexities, we implemented solutions aimed at enhancing user convenience, such as generating pseudo scores based on limited available information. This approach ensures a seamless user experience despite data variability. Moreover, our ongoing integration efforts with bank APIs prioritize interoperability and robust data extraction methodologies, aimed at overcoming format disparities and improving overall transactional transparency and efficiency.