A New Zealand insurer detected ongoing fraud that was highly sophisticated. The complexity in the way fraud was occurring made it difficult to detect with simple rules or manual inspection.

KPMG worked closely with the client to analyse several years’ worth of relevant claims data. The data was then cleaned, and integrated, and new attributes were created. The new data model was stored in SQL Server and made accessible via a bespoke Power BI dashboard.

Using these tools, our data scientists performed a series of assessments and drilldowns into anomalies and suspicious activity.

They worked with the client to further investigate claims identified as outliers that were not immediately identified as suspicious.

In addition to helping the client recognise fraudulent claims that it had previously not identified, KPMG’s digital tools have now enabled this insurer to spot a broader range of more sophisticated forms of fraud, on an on-going basis.