LogoLight logo
  • Who we are
  • What we do
  • Who we work with
  • What we think
  • Contact Us
LogoLight logo
  • Who we are
  • What we do
  • Who we work with
  • What we think
  • Contact Us
Latest Blog
Blog

AI-Driven Claims Processing

Investment in AI is one of the leading strategic objectives for organisations around the world. Boards of directors place analytics and AI as the No.1 and No.2 priorities – CIO Research, July 2020

The adoption of AI in business intelligence holds the promise of helping businesses improve effectiveness of marketing, perform segmentation to create personalised experiences, improve operational efficiencies and to aid decision-making. 

Today the use of AI is only in its infancy and real value is only being realised in a few instances, but this is about to change. 

By the end of 2024, 75% of enterprises will shift from piloting to operationalizing AI, driving a 5X increase in streaming data and analytics infrastructures – Gartner


AI
driven claims processing – a revelation 

Using AI to transform claims triage capabilities. 

At Bridge we are excited to share a recent design and deployment of an AI claims triage process for one of our clients.  One of Australia’s first implementations of AI driven claims triage. 

The project defined specific segments to support a tailored and personalised claims experience for each injured worker.  The segments were refined through AI modelling which provided more accurate segment management and improved insights for decision making. 


Project background – Reinventing claims capabilities
 

Our government insurer and care provider client provides workers compensation cover for ~326k employers and ~3.6M employees. ​ 

They were seeking to establish a world class claims triage capability, transforming from an adversarial model to one that empowers and offers choice to injured workers and employers. And, very importantly, one that merges systems onto a single platform.  


Project purpose – effective AI segmentation 

The aim 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 then assign them to case managers with the appropriate skillsets.  

Additionally, the triage capability had to be able to re-assess the level of care required for each claim, once additional information became available. 


Key Lessons Learned 

Through the design and delivery of the AI claims triage process we identified 6 key lessons learned that we would like to share.  Although these lessons are based on a claims case study they can be readily applied when using AI in any other business process. 


Accurate injury coding:
 ​

The type of injury is often the greatest factor in determining how long an injured worker will be off work​. Transformation to a more granular injury coding system resulted in a more specific injury classification, which in turn allowed for accurate segmentation.


Define clear segments:
 ​

You should determine the optimum number of segments for your claims and operating model. Plus define what a simple claim, ie a candidate for a straight-through process, looks like. This allows you to allot more time to complex claims.


Machine learning technique:
​

When a claim is lodged, the data points are often limited to mandatory fields. As the claim progresses through its lifecycle, more data points become available. Using different ML models at different ages of the claim can be beneficial to improving your triage accuracy.


Unstructured data sources:
​

The richest information associated with claims can be harnessed from your unstructured data. AI techniques, such as natural language processing, can help to identify risky claims where structured data can’t. You can extend NLP to voice recordings, though bear in mind the ethical implications of using voice analytics.


Business rules vs ML predictions:
​

Machine learning can increase the efficiency and effectiveness of your claims model, but it should be used to augment rather than dictate your actions. Once confidence in the ML models grows, your business rules can be relaxed.


Start small, build trust in the capability:
 ​

Start out with a rules-based approach integrated with a simple model, to help build trust in your machine learning. Once you’re confident, ML can take on a bigger role, with more sophisticated version


For more 
information please refer to our case study AI Driven claims processing – Lessons Learned, or give us a call and we’ll be happy to share our experiences with you. 


Contact Bridge Consulting

Bridge is a professional services firm specialising in data and analytics. We help our clients turn data into value through our dedicated focus on data strategy, data engineering, analytics, AI and data management. To discuss how we can establish AI-Driven Processing for your organisation, get in touch with our Bridge Consulting experts today.

View all posts by Clare Point

0
Share
Subscribe to our mailing list




Previous thumb
Previous Article
Data Validation and Assurance
Scroll


 

CONTACT US

Sydney +61 2 8599 3554

Melbourne +61 3 9900 6303

Email hello@bridge.consulting

 

FOLLOW US

Keep updated on all things Bridge

Bridge LinkedIn   Bridge Twitter

JOIN US

Careers

Privacy Policy

© 2019 Bridge Business Consulting Pty Ltd. All right reserved.