How analytics helped an auto insurer detect fraud

Author Information

wnsglobal has 37 Published Articles

India,
Maharshtra,
Mumbai,
Godrej & Boyce Complex,

How analytics helped an auto insurer detect fraud

Posted On : Mar-10-2011 | seen (732) times | Article Word Count : 643 |

Some months ago, a New York City court convicted 13 people including six medical professionals who had, over several years, milked auto insurance companies of millions in fraudulent claims. How did they do it - by staging ‘fake’ accidents in rental cars and then ‘treating’ the alleged victims to a battery of unnecessary tests including expensive EKGs and MRIs.

Auto fraud is clearly on the rise in the U.S. Consider this
• According to the Insurance Research Council (IRC), auto insurance fraud has added between USD 4.8 billion and USD 6.8 billion to auto claims in 2007
• Fraud claims lead to higher premia for honest customers, costing consumers an additional USD 200 - 300 in premiums alone according to Edmunds.com. According to the IRC, one third of all bodily injury claims for auto insurance contain some degree of fraud.

The concern is that detecting auto fraud, already a challenge, will only worsen for the following reasons

• Scarce resources - The underwriting resources necessary for detecting fraud claims are falling in numbers. A Deloitte Consulting report has forecast a likely shortfall of 84,000 in insurance claims workforce by 2014. It is believed that North America alone will experience a 40 to 50 percent shortfall in underwriting resources across all lines of the insurance business
• Overworked staff - Existing staff of investigators are already overburdened by the number of claims that they track. Therefore, their ability to manage a larger number of claims efficiently is a challenge
• Undetected fraud - Less than one percent of all claims are typically referred to for further investigation, implying that not enough number of claims are being scrutinized, and of that too, less than 0.5 percent of all claims are dismissed as fraudulent.
• Changing buying behavior - Increasing numbers of new auto policies are being purchased online. As the sales channels become more diversified, an insurance company's underwriting resources requires more expertise to track claims from a growing number of sales channels. The challenge is to detect and combat auto fraud in a cost effective and speedy manner. So let's see how investigators in insurance companies typically trace fraud and then ask whether or not these methods are relevant today. There are two traditional approaches for fraud detection
• Alerted to fraud through whistle blowers - Typically, this is someone in the know who informs a company of a fraud case which is often in the form of an anonymous call
• Using a rules-based approach - Such an approach attempts to derive execution instructions from a starting set of data based on ‘if this, do that’. For example, a computer-based system might help a doctor choose the correct diagnosis based on a cluster of symptoms. One way to trace likely fraud is by analyzing the information silo'd in companies on a person’s claims history, payment of premiums and residential addresses. By using computing power and statistical tools, it is possible to analyze information in order to throw up trends or profiles of potential fraud. Unlike traditional approaches, an approach that deploys analytics will provide Information on anomalous patterns that can help insurance companies investigate cases that would otherwise miss scrutiny Objective analyses of information that allows for the processing of larger volumes in a transparent manner.

As an example, WNS has worked closely with a leading auto insurance company that wanted to minimize fraud by understanding the profiles of fake customers, identifying patterns, and putting in place a proactive early warning system to support their investigation and fraud management teams. Using analytics, the company was able to improve the rate of genuine claims and significantly reduce costs associated with fraud claims investigation.

Business has become increasingly aware of the need to tackle fraud in cost-effective ways. For example, health insurance companies have taken a leaf out of the books of credit card companies and begun to analyze information that lies silo'd within their domains in order to be able to identify potential fraud cases, and tag them upfront. Auto insurance fraud costs genuine consumers, companies and the economy at large. Combating fraud with a well-thought out analytics plan will help insurance companies conduct their business efficiently, increase customer satisfaction and enhance their overall brand image.

Re-Publish this Article

<h1>How analytics helped an auto insurer detect fraud</h1> Author: <a href='http://www.articleseen.com/profile_wnsglobal.aspx'>wnsglobal</a> Auto fraud is clearly on the rise in the U.S. Consider this
•	According to the Insurance Research Council (IRC), auto insurance fraud has added between USD 4.8 billion and USD 6.8 billion to auto claims in 2007
•	Fraud claims lead to higher premia for honest customers, costing consumers an additional USD 200 - 300 in premiums alone according to Edmunds.com. According to the IRC, one third of all bodily injury claims for auto insurance contain some degree of fraud.

The concern is that detecting auto fraud, already a challenge, will only worsen for the following reasons

•	Scarce resources - The underwriting resources necessary for detecting fraud claims are falling in numbers. A Deloitte Consulting report has forecast a likely shortfall of 84,000 in insurance claims workforce by 2014. It is believed that North America alone will experience a 40 to 50 percent shortfall in underwriting resources across all lines of the insurance business
•	Overworked staff - Existing staff of investigators are already overburdened by the number of claims that they track. Therefore, their ability to manage a larger number of claims efficiently is a challenge
•	Undetected fraud - Less than one percent of all claims are typically referred to for further investigation, implying that not enough number of claims are being scrutinized, and of that too, less than 0.5 percent of all claims are dismissed as fraudulent.
•	Changing buying behavior - Increasing numbers of new auto policies are being purchased online. As the sales channels become more diversified, an insurance company's underwriting resources requires more expertise to track claims from a growing number of sales channels. The challenge is to detect and combat auto fraud in a cost effective and speedy manner. So let's see how investigators in insurance companies typically trace fraud and then ask whether or not these methods are relevant today. There are two traditional approaches for fraud detection
•	Alerted to fraud through whistle blowers - Typically, this is someone in the know who informs a company of a fraud case which is often in the form of an anonymous call
•	Using a rules-based approach - Such an approach attempts to derive execution instructions from a starting set of data based on ‘if this, do that’. For example, a computer-based system might help a doctor choose the correct diagnosis based on a cluster of symptoms. One way to trace likely fraud is by analyzing the information silo'd in companies on a person’s claims history, payment of premiums and residential addresses. By using computing power and statistical tools, it is possible to analyze information in order to throw up trends or profiles of potential fraud. Unlike traditional approaches, an approach that deploys analytics will provide Information on anomalous patterns that can help insurance companies investigate cases that would otherwise miss scrutiny Objective analyses of information that allows for the processing of larger volumes in a transparent manner.

As an example, WNS has worked closely with a leading auto insurance company that wanted to minimize fraud by understanding the profiles of fake customers, identifying patterns, and putting in place a proactive early warning system to support their investigation and fraud management teams. Using analytics, the company was able to improve the rate of genuine claims and significantly reduce costs associated with fraud claims investigation.

Business has become increasingly aware of the need to tackle fraud in cost-effective ways. For example, health insurance companies have taken a leaf out of the books of credit card companies and begun to analyze information that lies silo'd within their domains in order to be able to identify potential fraud cases, and tag them upfront. Auto insurance fraud costs genuine consumers, companies and the economy at large. Combating fraud with a well-thought out analytics plan will help insurance companies conduct their business efficiently, increase customer satisfaction and enhance their overall brand image.
About the Author: 
For more such article insights, please follow the link http://bit.ly/eMHhcl Article Source: <a href='http://www.articleseen.com/Article_How-analytics-helped-an-auto-insurer-detect-fraud_55335.aspx'> http://www.articleseen.com/Article_How analytics helped an auto insurer detect fraud_55335.aspx</a>

Article Source : http://www.articleseen.com/Article_How analytics helped an auto insurer detect fraud_55335.aspx

Author Resource :

For more such article insights, please follow the link http://bit.ly/eMHhcl

Keywords : analytics, auto insurer, detect fraud, insurance fraud , Scarce resources, Undetected fraud , claims, investigation,

Category : Business : Business

Print this Article Send to Friend