The power of analytics for insurance fraud detection: Opinion

Now insurance companies are increasingly relying on analytics to quantify and refine insurance fraud detection in a productive and efficient manner. Fraud or potential for fraud is being detected much earlier in the insurance cycle, with improved underwriting checks as part of the solution.
  • Updated On Apr 3, 2018 at 03:44 PM IST
By Kaustubh Deshpande

In the traditional method, insurance agents or loss assessors were assigned to fraud detections in insurance claims. They relied on the limited data and their intuition to detect a fraud and validate or reject the claim. They would typically use the available infrastructure to ascertain genuineness of documents, personally visit hospitals and sometimes solicit the help of specialists, doctors, lawyers or former policemen.

Now insurance companies are increasingly relying on analytics to quantify and refine insurance fraud detection in a productive and efficient manner. Fraud or potential for fraud is being detected much earlier in the insurance cycle, with improved underwriting checks as part of the solution.

Advt
However, for developing the capacity for fraud detection through analytics, it is important to take more of the available data and convert it into actionable intelligence to provide a broader view of the risk as experienced by the industry. Current industry standards are based on the broader assumptions about customer behaviour. In the face of accumulating losses and some negative customer perceptions around claims, it is important that consumer behaviour is studied as a whole and no longer as a reflection of what is visible through “on-us” data.

Sophistication in fraudulent practices
In many respects the insurance business is walking a thin line between identifying and preventing fraud and the need to pay out claims to instill trust in customers. Growing sophistication in individual or organised crime is a major concern for many insurance companies. According to The Economic Times report*, a major health insurer identified that an individual leading a fraud ring was found to have a whole organisation behind him, with reams of letterheads, rubber stamps, and other stationary items of various hospitals along with seven PAN cards with different names and 11 credit cards.

On the other hand, with claims repudiation rate at over 15% of registered health insurance claims in 2015-2016, according to official IRDAI figures, and over 15,000 rejected claims for life insurance in the same year, it is in everyone’s interest to address the issue of finding ‘true frauds’. A high repudiation rate or claims ‘written back’ can make business sense for an insurer on a case by case basis, but it also erodes consumer trust.

As the industry moves towards more digital underwriting and servicing, fraud detection and prevention efforts have to also adapt and become more data-focused rather than process-focused. In the first instance, bringing in the required data without interrupting the customer experience is necessary. Data that helps fulfill the promise of insurance, – which after all is the settling of genuine claims, but also able to identify and weed out fraudsters – is what insurers should be after.

Advt
On the organised crime front, it is going to require innovative and efficient tools to counter technologically-sophisticated fraud attempts. For this, it is imperative to develop a pooled database across the industry involving other stakeholders.

Industry is now demanding a new fraud detection and prevention model, one that involves multiple data analyses at the point of underwriting, policy renewals and periodic checks.

Role of analytics
While using multiple data sets definitely makes sense, the complexity of dealing with multi-faceted data quickly starts exponentially multiplying.
As data links from one source to other, it becomes really complex to find patterns that are emerging, in fact almost impossible for a human being to spot. Analytics will help insurers in sifting through a large amount of data and identify anomalies or fraudulent markers in claims, data patterns between insurers or across health networks and TPAs.

In several insurance markets around the world, we already use data linking technology that enables attaching the right data to the right person, and in developing a single customer view of the person and the risks involved. Once the data is linked, advanced analytics techniques allows to find the right patterns and bring them to the surface. It is almost impossible to imagine a big data solution today that does not use any analytical algorithm to make sense of the data.

For analytics to achieve its full potential, it is going to be important that insurers prepare for the data-driven practices necessary to get a full view of the insured person across underwriting, policy in-force management and claims. Usually the data flows in silos from various sources across various platforms. Combining this structured and unstructured data gives valuable insight into the history of the claimant.

Insights generated through data analytics can help insurance companies drive strategic initiatives, highlighting areas where claims can be paid more quickly, leading to efficiencies, and a clearer focus on complex claims (and those with a legal or regulatory aspect) that need more skilled handling.

Analytics: A method to improve the revenue model
The integration of a variety of data sources – for segmenting and propensity modelling – is a huge opportunity but also a concern for insurers.
Today some progressive insurers can pull data from various sources if an appropriate mechanism is in place. There are various sources like public records, Aadhaar, internal insurance company data, wider insurance pooled data, hospital registrations, other propriety data, news articles, structured/unstructured records, and so on. The issue that these insurers face is how to link data across multiple sources to make sense of it.

Data sources should be integrated step-by-step. When done correctly it means decreasing the volume of data and increasing content quality with every processing stage: integrating insight into the core insurance system, not just raw data. Insurers can then scan for fraud before a policy is underwritten or claim is approved.

The way I have described above is an evolutionary yet holistic approach to dealing with fraud, empowering insurers to monetize data from across the organisation, eliminating data discontinuity and segregating the appropriate data. Quality data available to the insurers hands them a competitive advantage in terms of risk management and helps improve the revenue model.

Analytics: A productivity booster

In summary, the power of analytics helps the insurer understand risks and trends associated with the latest fraudulent practices. This can generate powerful business insights that are crucial to improving underwriting precision and weed out fraud. Data-driven analytics can create an atmosphere which is conducive for growth and increased productivity.

The author of this article is Senior Director, Technology, India, Insurance – LexisNexis Risk Solutions.

DISCLAIMER: The views expressed are solely of the author and ETCISO.in does not necessarily subscribe to it. ETCISO.in shall not be responsible for any damage caused to any person/organisation directly or indirectly.
  • Published On Apr 3, 2018 at 03:44 PM IST
Be the first one to comment.
Comment Now

Join the community of 2M+ industry professionals

Subscribe to our newsletter to get latest insights & analysis.

Download ETCISO App

  • Get Realtime updates
  • Save your favourite articles
Scan to download App