Menu
search
Let's Talk
Back close

Business Impact

  • ~75%

    Effort Reduction

  • >95%

    Classification Accuracy & 90% Extraction Accuracy

  • 500+

    Entities Extracted per review

  • <160 secs

    To process document packets with >500 pages

Customer Key Facts

  • Industry : Insurance
  • Country : North America

Problem Context

  • The client wanted to automate the current manual policy review process
  • In the policy review process, the brokers compare old and new renewal policies to verify that existing coverage, limits, etc. are in place and data like deductible and location schedule are accurate
  • Currently, brokers manually go through hundreds of pages of documents of varying templates and formats to find the required data, this process is extremely time-consuming, inefficient and error-prone

Challenges

  • High error rate in manual review process
  • High operational cost since knowledge workers manually perform the comparison across multiple documents
  • No industry standardization of documents leading to over 5000+ different document formats and templates
  • Long timelines of up to 2 weeks leading to a loss in customers and low CSAT scores
Challenges

Solution

  • Dociphi, Quantiphi’s AI-powered Intelligent Document Management Platform assists the broker by automating the entire policy review process by eliminating the need for the broker to manually process documents
  • Dociphi seamlessly ingests and classifies numerous policy documents, quotes & binders of various formats and uses highly accurate ML models to extract data from multiple fields, key values, tables, hierarchical and conceptual data
  • Dociphi then compares the extracted date from the old policy, new policy and binder documents and highlights all the changes across all 3 documents, and provides the output in a format compatible with the downstream system

Result

  • Identification and classification of business cases & LOBs
  • Human-in-the-loop to maintain precision, ensure consistency
  • Bounding box to highlight extracted values
  • Improved PDF view and search facility
  • Extraction of tables, key values, hierarchical and conceptual data
  • Output format with downstream connectivity

Looking for similar project?

Let's Talk

Get your digital transformation started

Let's Talk