Big Data: The Key to Eradicating Insurance Fraud

By Jonathan Lewarne, Senior Director: Insurance Business Development at TransUnion

Quality information is the bedrock of the insurance industry – from the initial risk modelling that determines one’s premiums to the claims assessment process all the way to the policy renewal stage.

For local insurers, one of the most obvious advantages of Big Data lies in fraud reduction. Times of economic downturn tend to be accompanied by an increase in insurance fraud – wreaking havoc with insurers’ business models, and adding new layers of legal and investigative costs.

The broader populace of honest consumers also suffer from insurance fraud: in the form of hijacked policies, identity theft, and runaway increases in premiums (as they effectively subsidise the costs of insurance fraud). In fact, it’s a devastating trend that threatens to destabilise the entire insurance industry, a key pillar of the nation’s economic security.

By drawing on new, external sources of data, and using sophisticated algorithms to combine data-sets and detect patterns, insurers can go a long way to repudiating fraudulent claims.

There is a goldmine of insights available within the likes of credit bureaus, SA Insurance Industry Association data, the SA Insurance Crime Bureau, eNatis vehicle and driver databases, and other public databases.

So just how can Big Data be practically applied in the insurance business to identify suspicious claims while creating the fastest turnaround times for the majority of valid, honest customer claims?

  • Obtain and validate information earlier: When onboarding customers, one can cross-reference against 3rd-party databases and pre-populate information fields – drastically reducing the chances of error by the consumer, or the agent capturing the information. Known as ‘soft fraud’, it is at the point of policy creation (and not at the point of claim) that people often manipulate information – either the individual looking for a better rate, or the agent perhaps looking to meet certain sales targets.
  • Proactive fraud prevention: As the motto goes: prevention is better than cure. Insurers could bring down the standard fraudulent claim rate by an estimated 10% through the use of Big Data techniques at the onboarding phase – proactively preventing the writing of bad business into its books.
  • Predictive analysis: by layering data modelling onto various forms of customer data, exception reporting, financial trends, geospatial data, and even social media scanning, it becomes possible to predict which customers may have a higher propensity to commit fraud – leading to increased scrutiny for this group whenever a claim comes in.
  • Image recognition and voice biometrics: advanced verification technologies are now available, which help to ensure the legitimacy of the claimant through mechanisms such as voice-biometric systems in call centres), or the legitimacy of photo evidence or other documentation. Big Data is the foundation on which such technologies can operate, and should be woven into the Insurer’s operations.

A practical example of how Big Data could be leveraged is if a customer submits a vehicle insurance claim on a mobile device that is identified as being somewhere other than the car’s in-vehicle tracking system. In this instance a flag could be raised, and the claim sent for further data interrogation and investigation.

Ultimately, for insurers to truly take advantage of Big Data in claim management, they will need to build these sophisticated rules that funnel claims through the various phases of analysis and investigation. It’s critical that they understand the nature of the fraud committed against them, and tailor Big Data solutions that serve their most pressing anti-fraud needs.