Ten reasons to invest in an automated fraud detection system

Automated fraud detection systems are becoming more common in the insurance industry as the technology improves and the benefits become more evident.  Many companies have embraced this change and are showing measurable and significant returns on their investment. 

The quantifiable benefits are numerous – an increase in quality and quantity of cases referred to the SIU and an improvement in the impact rate on those cases.  But to truly understand the value of an analytical fraud detection system, companies should also consider the soft benefits even though they may be more difficult to quantify.

Here are ten of those benefits:

  1. Business intelligence - Being able to pull or automate reports from one system to monitor productivity, case diversification, strategic planning and goal setting, resource allocation, workload balancing and to help with regulatory reporting is an overlooked benefit.   Companies spend hours trying to pull information together and then manipulate it to get the type of reports needed. With a robust business analytics framework underlying your fraud detection system, it is easy to create, customize and standardize reports.
  2. Training - The ability to notify an adjuster, manager or investigator that certain data attributes triggered a high fraud risk score inherently trains the user. Seeing why claims are being flagged as high risk allows the adjuster to learn patterns and red flags that might be present on other claims.  This is especially important with newer employees who may not be focused on identifying suspicious claims as they learn the various aspects of their job.
  3. Identification of new trends, fraudsters and claim anomalies BEFORE payments go out - The right detection system will flag anomalous patterns such as high Bodily Injury to Physical Damage ratio in an auto accident or procedure codes not matching treatment plans in a casualty claim. While this analysis can be done on individual claims, a robust fraud detection system can also detect patterns of organized ring or provider fraud. For example, anomaly detection might flag a provider using a more severe diagnosis code than his peers for similar injuries. Adjusters and investigators would have to look through a lot of data and do a lot of manual manipulation to detect that trend.  With the right technology this can be done quickly behind the scenes.
  4. Level the playing field - In a manual process where the company relies on adjusters to identify suspicious claims, there are typically a few motivated adjusters who seem to make the majority of the referrals.  By having a system in place, you can ensure every claim is looked at the same way and find what those other adjusters have been missing.
  5. Improve customer service - By using the hybrid approach to identify suspicious claims and reduce false positives, suspicious claims get referred to the SIU or special handling group and the meritorious claims get paid faster or get to the right claim professional faster keeping the customers happier.
  6. Resource allocation, where are my clams happening and where is my staff located -   Do I need to shift resources to control expenses? Any strategically minded manager looking to use his resources efficiently should be asking these questions. Having the data to see this is powerful.
  7. Global/Enterprise level solution - Many companies have gone through mergers and acquisitions and have several systems or databanks that have various data sets in them, some have subsidiaries or business units around the country or globe for that matter.  Different business units all are trying to keep the profits of the company strong.  Having a truly global perspective of the anti-fraud efforts throughout the enterprise often leads to best practices and knowledge transfer.
  8. Data integration - Bringing in third-party data like public records that may have a predictive value is a big benefit. Data sources with  derogatory attributes like bankruptcies, liens, judgments, criminal records, foreclosures or even address change velocity to indicate transient behavior are all public records that can be integrated into a model. Other types of third-party data would be beneficial in enhancing efficiencies and could include appraisal information to determine if damages match description or loss or injuries being claimed. One of the most underutilized data source is the medical bill review data.  This data, if used in a model properly, is a gold mine for companies investigating medical fraud.  Uncovering anomalies in billing and adding these to the other scoring engines or Social Network Analysis will decrease the amount of time an investigator or analyst spends trying to pull all of the pieces together.
  9. Deterrence – When a company has a systematic approach to identify suspicious claims and then resists the non-meritorious claims, the ethically challenged fraudsters look for other sources of revenue and are going to travel the path of least resistance. This means going to a company that is not able to detect or investigate suspicious activity.
  10. Technology can be expanded to other areas - The technology used to identify suspicious claims can also be used to identify which cases have the highest likelihood of subrogation recovery or which claims can be fast tracked for quicker payment and better customer service.  It can also identify potential high severity claims early in the claim process.

Return on investment, enhanced efficiencies and quality referrals are important to express as benefits to a fraud detection system when deciding to upgrade or purchase a new process or system.  It is equally important to realize there are other benefits not usually discussed when looking at automated fraud detection systems.

I’m Dennis Toomey, Senior Solutions Architect in the Fraud and Financial Crimes Practice at SAS. For further discussions connect with me  LinkedIn or Twitter.

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>