Shift Claims Fraud Software

Most insurers have a huge repository of existing and historic claim and policy data in addition to a steady stream of new claim and application data. However, much of this data is stored separately across siloed divisions. Additionally, accessing external data sources, which can also help to identify fraudulent claims, is hampered by increasingly restrictive data privacy laws in Europe.

Shift is addressing these issues in two ways:

1) Proprietary Machine Learning Algorithm: Shift’s software is based on proprietary algorithms and recent breakthroughs in data modeling that allow Shift to cluster and parse unstructured claim and policy data sets from multiple sources.

2) Scale: The second factor that differentiates Shift is the ability to analyse terabytes of policy and claims data. Although competitors such as Guidewire offer fraud detection software, the ability to cluster and parse large data sets (which ultimately improves the accuracy of Shift’s claim detection software) from both internal and external data sources at scale has differentiated Shift.

The user experience for claims handlers is as follows:

1) Shift first highlights suspicious claims on a web interface along with a score and an explanation of the suspicion indicators. Visualization of suspicious claims is clear whether claims handlers are looking at individual alerts or suspicious networks.

2) Shift’s software then suggests next actions to a claims handler to investigate on a specific claim.

3) Although Shift’s software is able to detect both subtle and obvious signals of fraud (including group fraud), Shift’s software also integrates the expertise of insurers to interpret the validity of a claim. Specifically, Shift works with claim handlers, helping them use and understand Shift, until they are comfortable with all the features and functionality.

Shift has recently established an R&D partnership in Hong Kong in order to enable it to process Chinese characters in addition to enhancing its image recognition capabilities.

The Shift team are based in France but are focusing on Asia where higher growth rates combined with disproportionately high insurance fraud rates provide a suitable environment for Shift’s solutions.

Shift presents a transparent and streamlined experience for insurers:

• Shift is currently achieving a 75% accuracy rate on claims that it identifies as possibly fraudulent
• Enables insurers to lower premiums achieved through the reduction of fraudulent claims
• Operational cost savings achieved through the streamlining of loss adjustment and claims handling departments.

Ultimately, although there are numerous digital insurance startups focusing on online distribution and B2C operations, the innovation count is much lower on the claim side. This reflects the complex nature of loss adjustment and the process of claims fraud detection. Shift is a rare example of software dedicated to fraud detection that is also integrating machine learning algorithms into its solution.