Business Impacts
40%
In potential savings
80%
Model accuracy
12,000+
Documents processed per hour
Client Key Facts
- Location : North America
- Industry : Information Technology & Services
Problem Context
The client, a provider of artificial intelligence and machine learning to the Financial Services industry, wanted to develop an Virtual Assistant that enables automated classification and extraction of content from scanned financial documents. This was aimed at eliminating manual dependency in Business Process Outsourcing for reviewing documents.
Challenges
- Processing new loans is time-consuming due to the manual review of documents and involves over 400 different document types
- Volume and size of documents together with fluctuations make the traditional client server computing model cost prohibitive
- Unstandardized and disparate documents prevent the use of regular expressions for extraction
Technologies Used
AWS Lambda
AWS Step Functions
Amazon EC2
Amazon DynamoDB
Amazon Elasticsearch Service
AI Processing Assistant for Loan Document Processing
Solution
Quantiphi developed a Virtual Assistant that automated the classification and extraction of content from scanned financial documents ranging to about 400 different types; thereby removing manual dependency out of business process outsourcing.
Result
- Successfully digitized the processing of 12,000 pages per hour with an accuracy of 80 percent, leading to a significant reduction in human effort required for manual extraction of data
- Serverless architecture helped reduce infrastructure cost by 30 percent
