Automated With Artificial Intelligence

Making Document Classification Effortless

torrent-power
CASE STUDY

Document Management Service Powered
with Artificial Intelligence

Torrent Power, a Electricity firm started its search for a enterprise that can make its dream of managing and maintaining records easy and quick. After a quick understanding of client’s needs, Softage concluded to use Artificial Intelligence and Machine learning in order to initiate a better storage and quick retrieval perception for the client. So, Softage geared itself up with latest techniques including OCR, Opencv and Tesseract to get going with the project.

Interested in similar services from us?

To start, please complete our contact form and we will get in touch with you shortly.

Torrent Power

The Client

Torrent Power Limited is an India firm that is engaged in the generation of electric power, transmission and distribution. Currently the company is operating in different states of India including Gujarat, Maharashtra and Uttar Pradesh. The company has its firm stand in cities of Agra, Ahmedabad, Bhiwandi, Gandhinagar, Surat.

TASK OVERVIEW

Challenges

The major challenges faced for when the client purposed us the plan were identification of the documents and its classification. There are around three types of forms used and initial step was to make it identify using an automated system. After identification, process is moved on to separation of documents on the areas of interest and further complete scanning of the data present in form of record/document.

After analyzing the data on basis of areas of interest and types of forms, documents are moved on to the final level where the system uses algorithms of Artificial intelligence and Machine learning to classify the records.

Challenges
OUR PROCESS

Wire framing the ideas

Implementing AI and ML
After development of framework, our experience developers started to lay out different tools and technologies for AI and ML including OCR, Opencv and Tesseract.

Training Phase
If the feasibility is a success, we request a data volume of around 5,000 - 10,000 nos. for training. Afterwards, Our AI engineers will proceed with the training. Training will be an iterative process until the results are achieved.

Testing for Robust Output
Aggressive testing sessions were initiated to make the system robust and making sure that the client's expectations are met completely.

ai-process
Results

Torrent power was served with the best solutions that merged machine learning and document classification to provide the best in class document management system as required.

60%

Operational cost per month saved for client

20,000

Forms tag per day with machine learning

50%

Reduction in turn around time