A prominent Middle Eastern bank wanted to automate its financial spreading efforts to ease pressure on its analysts and lower operational costs. They needed on-premise software to support reading borrower reports in multiple languages, including Arabic.
Evalueserve’s Spreadsmart product was a natural fit for those needs, seamlessly integrating with their existing credit lending platform. Spreadsmart’s data extraction for the client was 99.6% accurate and 70% faster than manual efforts.
The Middle East bank wanted robust, sophisticated software to automate financial spreading. Financial spreading is the process of transferring data points from a borrower’s financial statements to a bank’s analysis program. Within the credit lifecycle, financial spreading is often a bottleneck.
The bank wanted to automate financial spreading because:
- Manual updates are prone to error, necessitating quality assurance analysts and increasing operational costs.
- Analysts were under immense pressure to meet deadlines quickly, with an expected 24- to 48-hour turnaround time.
- The bank’s financial spreading work was subject to surges, with increases of 200% between April and June every year, simultaneously increasing the possibility of manual errors and delays.
The bank chose Evalueserve’s Spreadsmart product because it:
- Is compatible with on-premise deployment, meeting confidentiality requirements.
- Can read and extract data from financial statements in Arabic, which is one of the bank’s primary business languages.
- Could seamlessly integrate with their existing credit system using APIs.
Evalueserve quickly configured and deployed Spreadsmart across all of the client’s credit teams, with over 150 licenses. Because the product uses a microservices architecture, it seamlessly integrates into the bank’s existing credit lending platform.
When a user uploads a financial statement into Spreadsmart, the program uses intelligent optical character recognition (OCR) to parse data and make it searchable even if scanned. It uses ML to identify and classify the table type and relevant information available in the notes. The software communicates directly with the bank’s credit system via APIs. The information extraction engine is powered by technologies including pattern recognition, sentence embedding, classification, customized NER, and deep learning.
There is no need to reconfigure anything to make the environment run well. By moving from manual financial spreading to automated financial spreading with Spreadsmart, the Middle Eastern bank achieved 70% faster data extraction. Spreadsmart and its built-in model validations led to 99.6% accuracy in data extraction.
Talk to One of Our Experts
Get in touch today to ﬁnd out about how Evalueserve can help you improve your processes, making you better, faster and more efﬁcient.