Digitization of Loss Run Underwriting Process

Insurance
case study

Business Impacts

~93%

Classification Accuracy

~90%

Extraction Accuracy

18

Fields extracted from loss summary, 18 from claims details

Customer Key Facts

  • Industry : Insurance
  • Country : United States
  • Size : 1000

Problem Context

The customer is a large commercial P&C insurer who wanted to automate and simplify the submission process, extract relevant data, and convert it into a standardized format to generate valuable insights.

Challenges

  • The manual time-consuming process of extracting data from long loss runs documents
  • Difficulty in processing the loss runs due to high variation across different carrier loss run documents
  • Limited insights from the extracted data

KPIs Tracked

Claims  by Nature of Injury

Claims by Nature of Injury

Claims by Class Code

Claims by Class Code

Claims by Month & Year

Claims by Month & Year

Claims by Claim Status

Claims by Claim Status

Total Incurred &  Total Paid

Total Incurred & Total Paid

Total Incurred by Year

Total Incurred by Year

Total Expenses  by Year

Total Expenses by Year

Total Paid by Year

Total Paid by Year

Solution

Quantiphi digitized and simplified the Loss Run Underwriting Process for the customer to extract relevant data. The extracted data was converted to a standardized format to generate actionable insights and make decision-making faster.

  • Identified 9+ distinct carrier templates that cover ~50% of the data and continue to add 2-3 carriers bi-weekly
  • 18 claim detail fields and 10 loss summary fields extracted
  • Input to the IDP through UI - the productionalized version to have a connector to Downstream Systems like Laserfiche
  • The output is provided in a standardized JSON format for generating valuable insights

Result

  • Automation of submission process and data extraction
  • Industry-leading classification and document extraction accuracy
  • Downstream integration of standardized output

Thank you for reaching out to us!

Our experts will be in touch with you shortly.

In the meantime, explore our insightful blogs and case studies.

Something went wrong!

Please try it again.

Share