Using data to fight insurance fraud
Insurers should consider a data warehouse solution
Insurance is one of the world's largest industries, with gross written premiums standing at approximately US$4.8 trillion (S$6.5 trillion) in 2017.
It is predicted to increase by 3 per cent to 4 per cent between 2017 and this year.
Perhaps it is due to its sheer size that criminals looking to profit from fraud are so attracted to this industry.
According to a Reinsurance Group of America (RGA) survey, 3 per cent to 4 per cent of all global claims are fraudulent, with the highest incidence (4.16 per cent) in the Asia Pacific.
Evidence suggests the reason fraud is taking place more in the Asia Pacific than elsewhere is likely due to a lack of proper oversight or sufficient regulation.
In an effort to overcome fraud as well as waste and abuse, many companies are turning to data and - more importantly - to an understanding of what the data means and how it can be used to protect themselves against criminal activity.
Just one insurance claim can involve dozens of data points that must be analysed and interpreted.
Multiply this situation across the insurance provider's complete book of business, and the data sets can become unmanageable.
Traditional on-premises cyber security and anti-fraud solutions simply cannot keep up with this volume as these tools are difficult to manage and nearly impossible to scale to the degree required.
The volume, variety and velocity of the data moving through the system is beyond its capabilities.
To solve this, insurers need a data warehouse solution.
These cloud-based solutions allow organisations to keep pace with expanding demand and also make it possible to accumulate, integrate and analyse unlimited quantities of data.
That frees up resources to automate processes such as risk and pricing analysis, fraud detection, provider abuse prevention, reimbursement analysis and litigation propensity scoring.
For example, property insurers use data analytics to detect and mitigate fraudulent claims.
Anticipating fraud without a predictive analytics platform is a rigorous, time-intensive process that depends on spreadsheets and number crunching.
With machine-learning powered by historical fraudulent claim data, insurers can "train" algorithms to proactively monitor and identify high risk claimants, which reduces manual effort and raises the effectiveness of fraud detection processes.
A data warehouse built for the cloud can enhance these predictive analytics capabilities while allowing companies to keep all business data in one place and process it faster, clone data without increasing storage and time travel to compare old data to new data.
Unlike on-premises systems that don't easily scale, a cloud-built data warehouse enables organisations to keep pace with the growing demand for insurance data. It delivers elasticity when scaling storage in the cloud so that the organisation's decision making can happen in a shorter period.
Data, for actions such as analysis or risk scoring, can also be shared in real time in a secure environment.
There are many reasons for insurance executives to harness insurance analytics, and combating insurance fraud should be the most compelling.
As insurance evolves into a more data-driven industry, the demand for modern, cloud-built solutions will continue to expand to support this growth.
The writer is vice-president, Asia Pacific and Japan, Snowflake Computing.