Investment Banking Cover Letter for a Large Financial Institution
95%
Accuracy
62
samples
$750K+
in annual savings
CHALLENGE Overview
Each year, the trade finance field is inundated with thousands of unstructured cover letters, containing critical information such as sending bank data and paying bank data.
Despite previous attempts to automate the extraction of data from semi-structured documents, the accuracy levels required remain elusive. Moreover, the current solution employed by the company falls short of providing the high accuracy necessary for optimal performance.
Client
Large Financial Institution
Goal
Streamlining error-prone, manual review of unstructured cover letters
WHAT WE DID
- By indexing and classifying documents at the field level, our model has successfully resolved the error issues that previously required manual review. With our innovative solution, you can now easily classify and extract data from cover letters, while exceptions are automatically routed to staff for further processing.
- With each manual intervention, AYR’s machine learning models become more complete and confident, thereby enhancing the platform’s intelligence and accuracy.
IMPACT
- 95% accuracy of extraction achieved in 24 hours
- Only 62 samples were needed to see these results, allowing you to achieve more while working with less samples
- $750K in annual savings thanks to the reduction in errors and increased efficiency/productivity
- Results in 24 hours
