Intelligent Automation – Using Computer Vision

For a Fund House.

Customer’s Objective

The MF forms in a leading fund house were scanned and sent across to be manually entered by employees with multiple levels of checking to ensure correctness of data. This used to take around 6 minutes per document to be filled on the web form incurring huge loss. The client wanted to reduce the time taken to manually populate the information.

Our Approach

We used the Computer Vision Techniques with a layer of Machine Learning to extract text from images and scanned documents and populate the relevant fields in a web based form and highlight ones which needed attention. The Engine could pick and complete forms and also validate against PAN and Aadhar number, bank details etc. Advanced quantitative approaches like Computer Vision, NLP and text similarity were used to solve the situation.

Business Impact

This helped in saving more than 1.5 Million Minutes per Month. Their workforce could now process increased volumes of 15% of monthly documents