USING ANALYTICS TO DETECT INSURANCE FRAUD - SAS · • ALFA estimate that fraud 15% of claims paid,...

Copyr i g ht © 2015, SAS Ins t i tu t e Inc . A l l r ights reser ve d .

USING ANALYTICS TO DETECT

INSURANCE FRAUD

SAS AZERBAIJAN ANALYTICS SUMMIT, FEB 3RD


AGENDA

Introductions

Customers

Solution Overview

A Day in the Life of A Claims Representative

Claims and Underwriting

Next Steps


AGENDA

Introductions

Customers

Solution Overview



Next Steps


GLOBAL INSURANCE CLAIMS FRAUD

CAIF estimate $80bn losses annually. III estimates $32bn for P&C alone

• Insurance Council of Australia estimates that between 10-15% of insurance claims exhibit elements of fraud

• SAICB estimate that 30% of short term insurance claims include fraud

• Schweizerischer Versicherungsverband estimate that 10% of claims paid are fraudulent

• GDV estimates that insurance fraud costs circa €4bn per annum

• IFB estimates that undetected fraud = £2.1bn adding about £50 to average premium

VvV (2013) estimate that fraud has increased 25% in last 5 years adding 150€ to a policy. Estimate that 10% of

claims may be fraudulent

• ALFA estimate that fraud 15% of claims paid, or 4-8% of premiums collected equating to €2.5bn per annum

• Svensk Försäkring estimate that 5-10% of claims include fraud, which is between 2.5 and 5bn Kroner per annum

• FFI 2014 survey - 19% said they knew a person “who has deceived his/her insurance company”.

• Malaysian government estimates insurance fraud to be RM1.74 billion. 30% of respondents to

survey thought it was OK to pad claims


SAS FRAUD

DETECTION &

PREVENTION

SOLUTION FOR INSURANCE

• The Solution offers an end-to-end framework which:

• decreases fraud losses - save up to 2% annual claims spend

• increases efficiency in the claims handling process

• The Solution is enriched with industry knowledge from SAS’ global insurance experience:

Single, integrated platform built on SAS

Full transactional, entity, product and network centric monitoring

Insurance specific solution – ‘white box’

Scalable to national level data volumes

Ability to work in real-time and in batch

Across different LOBs – personal & commercial

“Return of price was realized

within first 5 months of its

operation.”Maya Mašková, Anti-Fraud Coordinator,

Allianz, Czech Republic

Life

Medical

Disability

Motor

Property/Fire

Workers Compensation


FINDING THE FRAUD

100 MOTOR CLAIMS


FINDING THE FRAUD

100 MOTOR CLAIMS – TODAY THERE MAY BE 10 POTENTIAL FRAUDS CASES


FINDING THE FRAUD

100 MOTOR CLAIMS – MOST INSURERS CAN TYPICALLY FIND BETWEEN 0.25 AND 1 WITHOUT ANALYTICS


AGENDA

Introductions

Customers

Solution Overview



Next Steps


SAS FRAUD

FRAMEWORKINSURANCE SOLUTION CUSTOMERS

Plus others under Non-Disclosure Agreements

And very recently, Aksigorta…


"We have an excellent working relationship

with SAS. They took the time to learn from

us and truly understand the nuances of

claims fraud at CNA…”CZK 10m

Additional fraud detected in first 6 months

26%More cases investigated by same team

40%More proven fraud cases for criminal

prosecution

“The decision to

choose SAS

was a good one.

We’ve been able

to expand the

number and

type of insured

events that we

examine.

Maya Mašková,

Anti-Fraud

Coordinator





claims fraud at CNA…”$18mAdditional fraud detected in

first 2 years

$6.4mDetected by 4 predictive models

Tim Wolfe

Assistant VP

Special Investigations

Unit





claims fraud at CNA…”TL 1.8M

Additional fraud detection saving

«While the

insurance firms

are able to catch

only 5 percent of

all fraud through

their own means,

with analytics we

believe we can

catch as much as

25 percent of the

fraud.»

SBM CEO


CASE STUDY CESKA POJISTOVNA

Ceska Pojistovna is

• largest life and non-life insurer in Czech

Republic

• Part of the Generali group

SAS Fraud Framework for Insurance

• Using on motor, property and life books of

business

• 2014 saving estimated at CZK 20m

Benefits:

The Fraud Management System

paid back its costs as early as in

the first year since the launch of

the project. Česká Pojišťovna

appreciates mainly its innovative

approach to the evaluation of

suspicious damage claims, time

savings, higher efficiency and

simpler and more accurate work

with historically interconnected

data.


CASE STUDY ALM BRAND

Alm Brand is

• 4th largest insurer in Denmark

• GWP of €880m – 2013

SAS Fraud Framework for Insurance

• Using on motor & property books of

business

• 2013 saving $5.6m – 50% increase on previous

activity before SFFI

• 2014 saving $8.5m

“In order to protect the vast majority of

our honest clients against the

fraudulent practices of the few, we have

introduced SAS® Fraud Framework.

We always assume that all customers

are honest − but at the same time, we

know that we reveal fraud to the tune of

more than EUR 5.3 million annually. We

are more and more dependent on self-

service to let our clients file their

insurance claims at the time and place

which suit them best. To make sure that

we can still find the fraudsters even

when people are filing claims online, we

are dependent on strong processes

and systems,”

Brian Wahl Olsen, Director of

Claims.2013 story here.

http://sas-thinkforward.info/content/articles/01_2015/interview_alm_brand/


AGENDA

Introductions

Customers

Solution Overview



Next Steps


CHALLENGES DRIVING CHANGES

Fraud is on the rise & today’s

schemes are:

• Increasingly sophisticated –

organized, patient, share rules

• More agile, evolving strategies

• Higher velocity/faster

• Cross industry

• Multi channel

• Advanced technologies

• Engage insiders to understand

detection environment

Current Fraud Systems are no

longer fit for purpose

• Silo’d by line of business - No

sharing of data

• Act on transaction or customer

• Rules and predictive models

alone have limitations

• No real proactive steps taken

to combat cross channel fraud

• Evidence insufficient to act

upon

• Investigation time consuming

FRAUD &

FINANCIAL

CRIME


$€¥ VALUE 40%60%

Claims Exaggeration Deliberate Fraudsters Criminal Gangs

THE CLAIMS FRAUD LANDSCAPE


SAS FRAUD

FRAMEWORKSAS FRAUD FRAMEWORK FOR INSURANCE

Business Analytics Framework

Data Quality & Integration

Advanced PredictiveModeling

Business Intelligence

SFFI Core Components

Alert Generation

Alert Management

Network Analysis

Real-time Decisioning

Case Management

Insurance Lines of Business

Life & Health

Home General Liability

AutoWorkers Comp. Detection

Prevention

…


Analytic

Decisioning

EngineAutomated

Business Rules

Anomaly

Detection

Predictive

Modeling

Text

Mining

Database

Searches

Social Network Analysis

Known Patterns

Unknown Patterns

Unstructured Patterns

Known Fraud

Complex Patterns

Unexplained Relationships

BUSINESS

ANALYTICSA VARIETY OF TECHNIQUES


SAS FRAUD

FRAMEWORK

SOCIAL NETWORK ANALYSIS

Network scoring: Rule and analytic-

based

Analytic measures of association help

users know where to look in network

Net-CHAID for local area of interest

(node) in the network

Density, Beta-Index (network)

Risk ranking with hyper geometric

distribution, degree, closeness,

betweenness, eigenvector, clustering

coefficients (node)

Modularity (sub-network)


SAS FRAUD

FRAMEWORKHYBRIT SCORING TO FIND FRAUD

FR

AU

D

POPULATION

█ ANALYTIC

DECISIONING ENGINE

WITH SNA – HYBRIT

SCORE

█ ANALYTIC

DECISIONING ENGINE

WITHOUT SNA

█ RANDOM

I.e. if you examine 50% of the

population, you would expect

to find 50% of the fraud

Adding SNA variables can double the accuracy of your scoring models!


SAS FRAUD

FRAMEWORKREPORTING CAPABILITES


SAS FRAUD



SAS FRAUD



Claims Exaggeration Deliberate Fraudsters Criminal Gangs

BUSINESS

ANALYTICSDIFFERENT TECHNIQUES FOR DIFFERENT FRAUD TYPES

Insurance Rules

Database Searching

Anomaly Detection

Advanced Analytics

Social Network Analysis

Text Mining

Some differences for different LOBs


AGENDA

Introductions

Customers

Solution Overview



Next Steps


A DAY IN THE LIFE OF A CLAIMS HANDLER


A DAY IN THE LIFE OF A FRAUD INVESTIGATOR


A DAY IN THE LIFE OF CLAIMS MANAGEMENT


AGENDA

Introductions

Customers

Solution Overview



Next Steps


APPLICATION

FRAUDCHANGING TREND: MOVING TOWARDS PREVENTION

Challenge to reduce losses now bigger than ever

Sophistication of fraud techniques still increasing.

Increased use of new channels, e.g. online driving new fraud

modus operandi.

Existing efforts somewhat exhausted

Organisations have invested in claims fraud detection.

Making a positive impact but need to get inventive and find other

opportunities in the value chain to reduce losses

Looking at prevention

More pro-active thinking is shifting the focus to application fraud.

By stopping a fraudster at the point of policy application you can keep the

‘bad’ out of the organisation.

Improving the application fraud process

Organisations need to ‘know their customer’, which relies on better use of

data, often already in-house.

A holistic view of the organisation with appropriate analytics allows the insurer to

generate an accurate fraud risk assessment of each application.


TACKLING FRAUD

HOLISTICALLYTARGET OPERATING MODEL

Triggered

score

Reject Action

Initial Fraud

Triage

Reject Action

Automatic

continual scoring

Policy

ApplicationClaim at

FNOL

Triggered

score

Initial Fraud

Triage

Full Investigation

Policy

Acceptance

Claim

Acceptance

SAS Analytics Hub

Un

derw

riti

ng

Cla

ims

No

RiskFraud

Risk

Fraud

Suspicions

No Action

RequiredConfirmed

Fraud

No

Risk

No

RiskFraud

Risk

Fraud

Suspicions

No Action

RequiredConfirmed

Fraud

No

Risk

Updates


AGENDA

Introductions

Customers

Solution Overview



Next Steps


WHAT NEXT?

Arrange a free one day workshop to

• Review applicability to Baloise

• Examine which techniques would make sense

to start work with

• Outline high level business case

• Explore potential roadmap

Output will include a full report

Copyr i g ht © 2015, SAS Ins t i tu t e Inc . A l l r ights reser ve d . www.SAS.com

THANK YOU

http://www.sas.com/

http://www.sas.com/

Date post:	04-Aug-2018
Category:	Documents
Upload:	vudan
View:	247 times
Download:	0 times

USING ANALYTICS TO DETECT INSURANCE FRAUD - SAS · • ALFA estimate that fraud 15% of claims paid,...

Documents