In reality, only a small percent of insurance claims filed each year are fraudulent. However, that small percentage costs insurers big bucks. According to the Insurance Information Institute, property and casualty (P&C) insurance fraud amounts to about $32 billion each year. This requires insurers to be diligent about rooting it out.
Most companies have data sources that are not being used or fully analyzed. Many companies are also limited by their technologies and skill sets. As a result, these businesses are unable to quickly and efficiently identify risks such as fraud.
One Week Hackathon Uncovers Insights
With the right approach, expertise, and technology, companies can quickly uncover risk. This is critical in industries like insurance in which fraud must be recognized before a claims payment is made.
Teradata combines data science, analytic technologies, specialized consulting methods, and intellectual property to find data sources not being used, or find new ways to use existing data, and accelerates the time to gain new business value. That’s the approach we took when we worked with a P&C insurance company. We held a one week “hackathon” with four of our data scientists and business consultants, and six insurance company specialists. We looked at claims data that had not been used before and applied it to three previously identified business use cases.
We used multi-genre analytics, in which multiple techniques are intelligently brought together to gain insights. Techniques included text mining, graph, cluster, machine learning, and predictive analytics. After just a few days, we were able to identify how the data could show potential instances of fraud.
The hackathon is an example of a business outcome led, technology-enabled approach that can deliver high value results quickly. In this case, we had results in just one week.
Low Risk Engagement Offers Freedom to Innovate
Teradata business analytics consulting teams are using the same approach we took with the insurance company to deliver high impact outcomes to other industries. The approach, called the Rapid Analytic Consulting Engagement (RACE), allows us to engage a business by prioritizing use cases that deliver the most impact and align with the company’s strategic goals.
RACE eliminates the significant investment and several month or longer requirement to get answers that is common with other consultative approaches. One problem with these other approaches is that by the time the business sees the results, if decision makers aren’t happy with the outcome, there’s a feeling that it’s too late and too much is invested to change course and start over.
By contrast, the RACE is a low risk, low cost engagement. Unlike other consultancies, RACE gives analysts the freedom to fail, which can sometimes lead to new discoveries. We have the freedom to try approaches that have not been done before, give them time to develop, then analyze the results. We’ll know in just a few weeks if the approach will deliver value and is worth pursuing, or if we need to make a change.
Uncovering Networks Delivers New Business Value
When we worked with the insurance company, we were able to use a subset of data to determine claims that had a high likelihood of fraud. This allows the small team of fraud investigators to focus their limited resources on the cases that have a relationship to a person, organization, or other entity already known to be fraudulent. We also created an Art of Analytics picture called Fraud Invaders that offered a detailed visual of those relationships.
The large dots in the picture represent known instances of fraud. The small dots are claims that have not yet been investigated, but if we see a line connecting a small circle to a large one, we see a relationship that is worth investigating. This is just a suspicion of fraud at this point. People buy cars, change phone numbers, move to new homes, and do other activities that could result in a coincidental connection to a fraud activity, so it doesn’t mean a claim is fraudulent just because of a connection. However, focusing on suspicious claims is a good place to start.
This ability to identify relationships can also benefit other industries. For example, in the highly competitive telco industry, companies can use our advanced analytics to show networks of family, friends, and business contacts. Telcos can use that information to determine the “alphas,” who are at the center of networks. If alphas change providers, they influence others to do the same, so telcos can work to ensure these customers are not at risk of churn.
Likewise, identifying networks can help with preventative maintenance. Sometimes parts are likely to fail at the same time. Knowing this information, vehicle manufacturers can proactively replace all of those parts when one fails, which improves customer satisfaction and loyalty, and eliminates the need for multiple trips to the repair shop.
Analytic solutions help businesses uncover new insights for their data. What business outcomes could you achieve if you could get deeper insights from your data?
Watch a video with Christopher Hillman, to hear him explain Fraud Invaders in greater detail.
Christopher Hillman is a Principal Data Scientist in the International Advanced Analytics team at Teradata basedÊin London. He has over 20 years experience working with analytics across many industries including Retail, Finance, Telecoms and Manufacturing. Chris is involved in the pre-sale and start-up activities of Analytics projects helping customers to gain value from and understand Advanced Analytics and Machine Learning. He has spoken on Data Science and analytics at Teradata events such as Universe and Partners and also industry events such as Strata, Hadoop World, Flink Forward and IEEE Big data conferences. Currently Chris is also studying part-time for a PhD in Data Science at the University of Dundee applying Big Data analytics to the data produced from experimentation into the Human Proteome.