Insurance companies have long used tried and tested techniques for detecting insurance fraud, but now high performance analytics enables them to be more proactive in identifying fraud before it occurs, said Tracy Dunbar, director of analytics for South African company BITanium.
Speaking to HumanIPO, Dunbar said insurers should ask themselves: “How can 20 or so generic rules, that are seldom updated, possibly target the complex and changing patterns of suspicious behaviour that people are carrying out on a daily basis?”
Dunbar said fraudulent claims have been a challenge for insurance companies for a long time and current estimates claim approximately 10 per cent of short-term insurance claims are fraudulent and fewer than 20 per cent of those are actually detected, which calculates to millions of rands being lost annually.
Dunbar explained the traditional techniques used in insurance fraud detection, which involves relying on a “combination of rules that are hard coded into the IT system and on the gut feel [of] claims assessors”.
Aside from the rules, which are seldom reviewed and updated, being managed by the IT department, they also do not target new and changing patterns of fraud.
“Fraudsters learn the rules and find loopholes.”
Furthermore, Dunbar said there is no business agility, which is due to the substantial requirement in requests and time to implement modifications and qualified assessors are in short supply and are usually expensive.
“Advanced analytics uses a combination of statistical and mathematical techniques and computing power to find patterns in data that are not evident using more traditional methods. Advanced analytics empowers insurers to develop models to target changing profiles of fraud and to implement different models for each line of business,” said Dunbar.
Dunbar said advanced analytics can be “implemented into real time solutions” together with rules to target low risk claims being paid immediately, complex claims sent to more experienced assessors and suspicious claims, which are investigated within the forensics department.
Regarding software tools providing advanced analytics, Dunbar said: “There are a number of software tools on the market which provide users with the relevant statistical and mathematical algorithms necessary to develop advanced analytics models.
“The leaders within this market are considered to be IBM SPSS (International Business Machines – Statistical Product and Service Solutions), and SAS (Statistical Analysis Systems).”
Furthermore, Dunbar said data provided by third parties or unstructured sources, such as social media platforms, can also be a valuable source in identifying fraud.
“Third party data can significantly contribute to the richness of the data, which in turn can increase the accuracy of the models. Data such as credit ratings and telematics data is extremely useful in the process of identifying suspicious claiming behavior,” said Dunbar.
In conclusion, Dunbar said high performance analytics is not simply “another technology fad,” because she believes it is a revolutionary tool, which delivers a measurable return on investment (ROI) as well as improving customer satisfaction and competitiveness.