Jeroen MORRENHOF, CEO, FRISS
Within Europe one out of every ten claims is fraudulent. In certain regions, such as CEE including SEE and CIS, this threat is at least three times the European average. So what does this mean for the insurance companies? In a word: loss.
XPRIMM: Is fraud a threat for the insurance business? Which are its effects in this field?
Jeroen MORRENHOF:
Yes, fraud is a real threat to the insurance business and throughout
the insurance processes including both claims and underwriting. For
claims, we know that within Europe one out of every ten claims is
fraudulent. In certain regions, such as CEE including SEE and CIS, this
threat is at least three times the European average. So what does this
mean for the insurance companies? In a word: loss. This loss then is
passed on to customers resulting in higher premiums. This cycle affects
us all. Considering that the average European claim is between EUR 1,000
and EUR 2,000, you can understand the extent of the problem and the
cost to the industry. Moreover, it is also is a social responsibility of
insurer to make sure that the honest clients don't suffer from the
fraudsters in your portfolio. It has been estimated that an average
European family pays an extra EUR 150 per year in premium to cover for
the fraudulent claims.
Manual detection of fraud does help but
technology is needed to enable a more cost-effective and efficient
process for the companies in detecting possible fraud cases. Organized
fraudsters are able to operate across various insurance companies and
across country boarders. New, better fraud detection methods must be
introduced to alert companies of possible fraud before the claims are
paid out or the new policy is written. Preventative risk management is
needed.
XPRIMM: What tools can insurers use to combat fraud?
Jeroen MORRENHOF:
The tools available today to detect fraud in insurance consist of
predictive models (based on historical data), behavioral models (used to
detect variations from normal behavior patterns), expert models (rule
engines with if-then-else type rules ) and social link analysis
(determines if network links exist between customers, third-parties,
objects, phone numbers, bank accounts, address, etc). The best model, or
hybrid model, is comprised of all four types of models to produce the
most accurate detection, the use of which is seven to nine times more
accurate than manual fraud detection. Also, the addition of external
data sources to the solution, such as social demographic, credit
scoring, government statistics, company registrations, for example,
provide even greater insight into detection. The more data is analyzed,
the more accurate the results.
And FRISS is continuously adding
new techniques to the platform. At the beginning of this year we
announced a partnership with Iovation (www.iovation.com)
for screening of the devices on a global scale. Also are we doing
pilots with photo screening technology to check to see when a photograph
was taken and if a photograph was altered in any way. This technology
will help a great deal in claims processing. Also, new technology will
soon be used to check online claims entries to verify if the individual
is in fact the policy holder and will analyze the stress in the voice to
detect possible fraudulent claims.
 XPRIMM: Is communication between insurers important to fight fraud?
Leetha SPYROPOULOU:
Communication is the key in fighting fraud. It is of utmost importance
for insurance companies, claims managers and underwriting managers to
communicate what is happening in their business across the departments
and share data and the issues they face surrounding the detection of
fraud.
On a country level, all insurance companies in a
particular market will benefit by sharing a shared claims database and
an incident database. Companies would only see hits on the fraud
indicators and not be able to see all data so there is no issue for data
protection. Thus, all member companies could protect themselves by
sharing accumulated data, as the more data the better the fraud
detection for everyone. Also, the companies are better protected from
fraudsters moving from one company to another.
FRISS is working
on a platform to take this even a step further and set up a system for
cross border claims. This means that, in the future, moving to another
country is being detected.
XPRIMM: Are there any visible trends throughout Europe regarding the fight against insurance fraud?
Leetha SPYROPOULOU:
With the recent economic crisis in Europe and the recognition in the
insurance industry of the problem of undetected fraud, Europe is
proactively taking action to combat fraud, on country levels and in
insurance companies.
In the past year alone, FRISS has opened the
CEE Offices in Greece and the DACH office in Germany, and soon to open a
CIS office in the near future, just to meet the demand and to indicate
the extent of the interest in fraud detection in Europe.
This
shows the increased interest of insurers across Europe to not only
discuss the topic but actually prevent and detect insurance fraud.
XPRIMM: Which markets are the most exposed to insurance fraud?
Jeroen MORRENHOF: Fraud
is everywhere, and throughout Europe, both in developed countries and
emerging markets. It is not a fair statement to say that fraud exists in
certain markets or regions. We must take many factors into account.
Generally, the greatest presence of undetected fraud exits in the CEE,
SEE and CIS regions where fewer controls are in place. That is why the
great interest of insurers in the region to really adopt a fraud
detection policy and implement a solution to protect their business and
their profitability.
XPRIMM: Does the development of online channels pose any threats for the company or for the customers?
Jeroen MORRENHOF:
Introducing online channels is a big project for many insurers. But the
focus of these projects is mainly marketing and technology. What we
have seen at many insurers it that the introduction of a new online
channel attracts a whole different group of clients. Clients that are
out there to do "market research" and find the weak spots in the
defenses of the insurer. But not alone fraudsters/organized crime is
targeting these new online channels. The online channel attracts a whole
new type of client. A type of client that's a different risk. And one
that is not captured in the historical data of the insurers. Therefore
we see very often that the pricing turns out to be incorrect and also
the tendency to switch much easier to another insurer making the total
cost of sales per policy high.
XPRIMM: Does a universal profile of the insurance fraudster even exist?
Leetha SPYROPOULOU:
No, unfortunately not. Fraudsters are constantly changing their
patterns in their behavior (modus operandi). That is why we need to make
use of fraud detection solutions. The more data sources we can check
against, the more accurate the fraud detection. We can, therefore, find
patterns of data and network links which help us to better identify
possible fraudsters.
Also, the system needs to learn. It's not
enough putting a system in place and configuring it with the latest
insights. They will become outdated. So it's permanent monitoring,
learning and tuning. When selecting a solution, make sure you can not
only learn but also implement the findings yourself and very fast,
because the fraud cycle continues.
XPRIMM: How can you determine if a fraud is imminent?
Leetha SPYROPOULOU:
Using the FRISS solution, a user is able to conduct an online,
real-time or batch screening, from that screening a FRISS score is
provided. This FRISS score is checked against an internal score
threshold which indicates the action needed to be taken. For example, a
RED score will indicate immediate action to be taken by a senior
official and provide indication on which fraud indicators were hit, or
where the fraud is coming from, such as a suspicious phone number
connected with a previous fraud case. The senior officer can then act
quickly to check the alert and to make recommendations to stop any claim
payment or to continue with legal proceedings, for example. The FRISS
solution has over 600 standard fraud checks built into the solution,
which can be added to and changed if necessary.
XPRIMM: What tools does your company provide to fight insurance fraud?
Leetha SPYROPOULOU: We
provide a dedicated insurance fraud prevention platform. All
applications run on the FRISS Platform. These applications can be used
separately, but when combined they form the most complete, effective and
intuitive solution available to manage fraud, risk and compliance.
During
fraud and risk detection, investigation, analysis and visualization,
new insights and knowledge are gained. The FRISS VALUE CYCLE© takes this
new intelligence and uses it to improve expert rules, risk profiles,
text mining, network detection and predictive models. The entire process
from underwriting to claims will automatically get smarter while using
the FRISS Platform.
The FRISS Platform is connected to numerous
databases and can extract data from a wide array of internal and
external sources. The main applications on the FRISS Platform work along
the entire insurance process.
The FRISS solution can be used by
any insurance company, small or large. The FRISS solution can be used as
a SaaS (software as a service) or implemented on-premises, according to
the data protection laws of the country in question and the preference
of the company. The screening through the detection engine can be done
in batch, online or via web service.
Editor: Mihai CRACEA
| Published on 04.06.2014 |
