Page 25 - Turkinsurance Digital Magazine

23

We use an analytical model to detect abuses the Gendarmerie and Coast Guard Acade-
and frauds. Every day, we share suspicious my. Our trainings also include legal and ma-
claim files with insurance companies and jor insurance modules.
warn them. We request them to examine
the files and give us feedback. We make de- What are the benefits of your system for
tailed examinations and analyses in many the insurance market?
points; such as looking for relations between
2 drivers who get involved in an accident. In At the beginning of 2017, we have shared
light of such data, we give scores to claims 420 million TL worth of suspicious files that
files. Our system is a continuously learn- were caught by our system in the auto insur-
ing analytical model. As it learns more, it ance line with insurance companies. Among
can present more accurate results. We cre- those files, cases totaling 24 million 675
ate relation scores using social media. We thousand TL in worth were approved by in-
can detect people and businesses that bear surance companies as abuse and fraud.
high risk of fraud and abuse. Moreover, we
conduct behavioral analyses which include In addition, we ask insurance companies to
several analytical projections such as their send us files about fraudulent and abusive
longevity in the system, total amount of acts that they have detecting using their own
their claims, regularity of their payments, means and add them to our system so that
and whether they were previously included it can continue learning. Since 2011, our
in risky files etc. system caught 32 thousand hits in terms of
suspicious people. And we have received 21
Founded with the aim of detecting abuses thousand reports. During 2017, we have
and frauds in the system, our reporting pool provided the insurance market with a ben- claims in traffic insurance. We are working
is a platform where everyone including citi- efit of 100 million TL by detecting frauds towards addressing that area as well. We try
zens, agents, companies or other institutions and abuses. This amount does not include to provide services for the insurance market
can report suspicious transactions. For in- the frauds and abuses detected by insurance with similar models in every branch.
stance, when an insurance company detects companies detected through their own sys-
a suspicious transaction or an actual abuse, tems and operations. We inform insurance What do you see when you compare Tur-
we enter this information into our system, companies about the abusive and fraudu-
which helps other companies take preventive lent cases and what they should be vigilant key to other countries in terms of insur-
measures against the parties of that event. about. The smallest word in a claims file can ance fraud and abuse?
However, we do not take action for abuses cause us to examine that file as suspicious
and frauds detected by an insurance compa- and sometimes even reject it. We examine the practices conducted in Eu-
ny using their own data. We generally focus rope and the rest of the world in terms of in-
on more organized abuse and fraud cases What are SİSEB’s current targets? surance abuse and fraud. We know that we
that are over certain amounts and concern still have a long way to go. Insurance com-
more than one companies. We continue to work for other branches panies need to notice and fight against in-
and conduct projects. We also try to in- surance abuse and fraud. We see that fraud-
We know that training is very significant form people and organizations about fight ulent and abusive insurance transactions are
in terms of insurance fraud and abuse. against insurance fraud on our website. We all around the world. For instance, we know
that 80 billion USD of total insured losses
Can you inform us about SİSEB’s activ- work integrated with medical institutions in in USA are fraudulent and abusive claims.
ities in this area? health insurance branch. Rate of abuse and British state that they catch 2.1 billion GBP
fraud is really high in this branch; howev- of insurance fraud and abuse. German esti-
We try to improve the perception of fighting er, as personal data is concerned and we mate that their loss is higher than 4 billion
against insurance fraud and abuse through are very sensitive about it, we use differ- EUR in this area. We estimate this amount
trainings we organize in cooperation with ent scoring methods in this area. We never has reached 1 billion TL in Turkey. 15 per-
Turkish Insurance Institute (TSEV). Every share personal information. Collecting and cent of paid claims in France and 10 percent
year, we organize 8 to 10 trainings in this safeguarding the information is of vital im- in the Netherlands are abusive and fraudu-
field. These trainings focus on a wide range portance; it is also significant how to serve lent claims. This ratio changes between 10
of audience from external investigators to that information. We are very sensitive in and 30 percent in Turkey.
insurance company staff. Five times a year, this regard. Physical damage is one of the
we also organize trainings that are led by areas that see highest amount of abusive
Zeynep Tütüncü / zeynept@bestyayincilik.com