Panos Kouvalis is Direct Business Manager at the Greek insurance company INTERAMERICAN, which is a member of Achmea. This company has around 600,000 motor vehicles in its portfolio, particularly via the direct channel ‘Anytime’. It’s one of the largest companies in this field in Greece.
Automated risk assessment
The identification of high-risk prospective insured persons and the prevention and detection of fraud is of a high priority. Thanks to automated risk assessment, identification takes place at the front door. This process increasingly uses information from the claims department. But the most important condition for success is a culture focused on internal cooperation and sharing knowledge.
We have been using enriched data for risk assessment for a little less than a year. Additionally, information from the claims department is now integrated into the acceptance system. The introduction of this method obviously requires time and money. But this is clearly positive since we have already earned back a major part of the investment with the cost savings afforded by this method.
We have a large motor vehicle portfolio, but only a small underwriting team. It is therefore very important that we have the best information available. And that the system automatically produces the desired signals based on this information.
By collecting and using as much information about the vehicle and the policyholder as possible when assessing the application, we can be more effective in risk assessment. This allows us to improve our prevention of fraudsters being accepted. It also allows us to clarify which newly insured persons should be given extra attention, because of a greater than average likelihood of damage claims; increased damage claims mean higher costs for us and all premium payers. This does not involve fraud and it usually does not mean that the prospective insured person is refused insurance. It may lead to extra eligibility conditions or a higher premium. By doing this the increased risk is not passed on to the average insured person.
At the moment, based on the data analysis and smart knowledge rules, the system filters-out approximately 3,5% of the new applications. These are then examined by our acceptors in more detail. About 10% of these filtered-out applications require extra measures in order to proceed to acceptance. Measures such as adjusting the policy offering, lowering the insured value and increasing the deductible. If fraud is involved, then the policy can be cancelled.
Obviously, we want that risk assessment become even more efficient in the future. Our goal is to decrease to number of false positives. As the system is operational for longer and more data becomes available, this will certainly be the case.
Incidentally, we do not only use the system for new policies. Existing policies will also be screened in this manner upon renewal and adjusted accordingly on the basis of the result. This means that in due course our entire portfolio will be checked in this way. And that each renewal will be rechecked to confirm that all signals are still green. So, this means an ongoing risk assessment strategy.
New system requires culture adjustment
To realize this, we have had to let go the old system of separate underwriting and claims departments. We add information from external sources to the assessment of the application. For example, information about previous – unclaimed – damage or the specified value of the vehicle and the accessories. In addition, there is also a check with regard to the accuracy of the data about the vehicle and the prospective insured person. We also add information collected by the claims department. Such as previous claims or fraud attempts or relationships with known fraudsters. By analyzing all this information, we get a good picture of the prospective insured person and his vehicle. Finally, we can make a responsible decision during underwriting via automated risk assessment.
Where previously the underwriting and claims departments operated separately, they now have to cooperate with each other. Employees must learn from each other, study the casuistry together and jointly examine which information is available and useful; become real colleagues and enjoy success together.
Cooperation leads to satisfied customers and better results
The need for cooperation is not limited to these two departments. Risk and fraud awareness must become embedded in the fabric of the company: from underwriting to legal & compliance, from claims to product development & sales. Naturally this includes the Top Management. Without the involvement and inspiration of senior executives, who must weigh-up interests responsibly, every change is a dead end.
With our modern data system and our collaborative culture, we will achieve a balanced insurance portfolio and have a good insight into the risks we carry. I am convinced that this will lead to better products, satisfied customers and better operating results.