Project Description

AI-Driven Claims Processing – Lessons Learned

Project Description

Our client was a government insurer and care provider. They provide workers compensation cover for ~326k employers and ~3.6M employees. They were seeking to establish a world class claims triage capability while transforming the claims operating model and merging systems onto a single platform. Our role was to implement a real-time analytics engine capable of segmenting >60k workers compensation claims per year based on their required level of care and assign them to case managers with the appropriate skillset.

What We Did

  • Scoped and planned for project, spanning 4 areas of the business (triage, claims, digital and operations)
  • Developed end-to-end solution architecture
  • Executed Claims Triage Engine architecture with Claims Management System, Front-end Web Portal and CRM
  • Transformed the claims model to utilise a granular international injury coding system
  • Created natural language for common injuries for front end users to access upon claim lodgement
  • Liaised with business and data scientists to assist in development of machine learning (ML) models
  • Developed and deployed business rules and their interactions with ML models within the claims triage engine
  • Provided quality assurance, implementation planning and post-implementation support
  • Ownership and management of program budget, implementation plan with business prioritisation and resources

Outcomes delivered

  • Successful delivery of productionised real-time claims triage analytics engine and integration within the new claims operating model
  • Ability to consume and analyse ~100 data points to recommend a predicted segment for each claim in real time for the duration of the claims’ lifecycle, improving claims handling decision making and efficiency
  • Ability to meet peak demands by processing 10 triage events per second. Approximately 500k triage events were processed per month.
  • Retention of transactional data for each triage event to enable retraining of the model for the purpose of improving triage accuracy
  • Ability to report on claims and insurance policies to track trends and accuracy

Project Details

  • IndustryInsurance
  • Service LineAnalytics + AI