Intelligent Document Processing

Finance

Business Impact

Significant reduction in time for processing the documents & upto 50% cost saving.

Customer Facts

Location: India

Industry: Finance

Problem Statement

Problems we Acquired

The sales team was manually scanning & proofreading the documents for approval of the loan. It was a very time-consuming process with high chances of human error.

Challenges

Obstacles we faced

Bad scan quality and stamp on the document
Low resolution images and faded old documents was an issue
Handwritten data and regional language made it difficult for processing

Technologies

Tech stack we used

Python

Pytorch

SQLite

OCR

Solution

How we made end solution

Collection of all the scanned documents for model training
Extraction of information using the OCR solution
The extracted information was processed in NLP model which understands and categorizes the data
The processed data was displayed on the integrated dashboard and updated in the system