Improve Underwriting Efficiency through Intelligent Content Extraction Solution

About the customer

The client is an A rated global provider of property, casualty, professional liability, and specialty insurance and reinsurance.

Background

The client receives 300-400 requests for quotes in day and these were received in unstructured, semistructured and structured formats over emails and attachments e.g. Word documents, PDF, excel or as email body.

The existing process was inefficient as the underwriters were spending time to manually review the enquiries as the inputs varied in formats which prolongs the submission cycle. The business was suffering due to loss in productivity eventually impacting the total business written. The manual process was error prone requiring rework.

Save for Later Download Case Study

Solution Highlights

To address the problem, Coforge deployed its proprietary framework solution accelerator named SLICE (Self Learning Intelligent Content Extractor) offering advanced technological capabilities to extract and ingest data from various structured, semi-structured and unstructured sources.

SLICE’s Computer vision, Machine Learning and Natural Language Processing capabilities for content extraction came in handy to extract meaningful information from 50+ email sources which included SOVs and slip details. These information patterns were then used to train the model deployed and was setup to improve performance combining the feedback received from underwriters.

The AI engine would also learn to make recommendations based on historical data apart from the ruleset output.

The architecture offered underwriters the flexibility to override system recommendations, where necessary, which can be used for self-learning for the system for future decisions.

The system was deployed for the Political risks line of business to start but was modular enough to scale for other Lines of Business with optimized effort.

Value Delivered

The automated data collection capability of SLICE improved underwriter productivity by 10-20% This was achieved due to a number of different features:

  • Near real-time extraction of content as soon as an email arrives to the recipient mailbox
  • Automated verification of insurance documents
  • Customised workflow based on the type of document received

The solution was developed as a web-based application that can be accessed as a standalone system or from Outlook. This delivered a seamless user experience for the business.