Reduction in end-to-end spreading time
Faster data extraction
Accuracy
Summary
The Wholesale Credit and Lending Team at a top global bank faced inefficiencies and errors in their lending operations due to manual spreading. The bank partnered with Evalueserve to automate the spreading process and address these challenges. With the implementation of AI-powered Spreadsmart, we helped the client accelerate financial spreading and eliminate human errors for faster, more accurate lending decisions.
The Challenge
In the turbulent lending environment of today, banks need to ensure credit operations are efficient as possible while maintaining accuracy to mitigate severe risks. Financial spreading is one of the most tedious, time-consuming tasks in the credit underwriting process. The bank’s financial spreading process was entirely manual and thus extremely inefficient and error prone. The team was facing several challenges, including processing financial reports in different formats, managing diverse accounting standards and languages, addressing inconsistencies in line-item treatment and exceptions, and the need for multiple reviews and quality checks.
Our Solution
Evalueserve partnered with the bank to undergo a complete transformation of their spreading process in three phases.
Phase 1
The initial phase involved implementing Spreadsmart, our proprietary AI-powered spreading tool. Spreadsmart was rolled out to over 120 team members to automate financial data extraction and deliver essential features such as:
- Auto table detection
- Robust and intelligent OCR capability
- Autofill mapping with multilevel taxonomy
- Built-in validations for reporting changes or document errors.
Phase 2
During the second phase of the automation, Evalueserve integrated Spreadsmart with the bank’s existing credit decisioning platform, Moody’s CreditLens. The seamless integration enabled users to search for CreditLens customers in Spreadsmart to synchronize data, access past spreads for analyst reference, and conduct and validate spreads in Spreadsmart before submitting to CreditLens with an auto pre-check mechanism.
Phase 3
In the third phase of the client's lending transformation, Evalueserve incorporated new AI capabilities into Spreadsmart for enhanced efficiency and reduced manual intervention, which accounted for around 18% of the spreading processing time.
The effort centered on three primary use cases to streamline operations and elevate the user experience.
- Table Recognition: Spreadsmart's AI recognizes tables within financial sectors, providing a user-friendly experience that can be adapted through additional training.
- Document Categorization: AI organizes data in financial reports, optimizing the classification process.
- Text Analysis: AI analyzes unstructured text passages, aligning data with bank guidelines and utilizing historical customer financial information for automatic input.
Evalueserve’s lending experts worked together with our AI/ML (Machine Learning) Center of Excellence to implement these use cases and integrate AI/ML algorithms into Spreadsmart.
Business Impact
Employing Spreadsmart resulted in significant achievements for the client’s lending operations, including a 70% acceleration in extracting data from financial statements with a remarkable accuracy rate of 99.6%. This implementation streamlined auditing and navigation for end-users and enabled the extraction of financial data from documents in multiple languages and formats. Moreover, it facilitated efficient governance reporting through real-time status checks and the generation of SLA reports.
With the integration of CreditLens in the ecosystem, the client experienced a 21% improvement in the end-to-end processing time for financial spreading.
The incorporation of AI into Spreadsmart increased efficiency by an additional 9% and significantly improved user interaction.