Microsoft PowerPoint - CFO_Jacek Sta\361czyk_14.35.pptx

Microsoft PowerPoint - CFO_Jacek Sta\361czyk_14.35.pptx

IBM Counter-Fraud
Management: Przyszłość
Integrated Fraud
Intelligence
Jacek Stańczyk
i2 / Counter Fraud
Organizacje tracą około 7% rocznych przychodów ze względu na oszustwa...
$994 miliardy tylko w US
Ekonomiczne spowolnie prowdzi do większych
oszustw
Jednostki i firmy szukają nowych sposobów na
zwiększenie przychodów, niekoniecznie legalnych
Warunki rynkowe wytwarzają presję na kientów
Potrzeba na znalezienie dodatkowych sposobów
na lepsze wyniki
Oszustwa stają się coraz bardziej złożone
Oszuści stają się coraz bardziej wyrafinowani
Postępy w analityce i Big Data pozwalają na
bardziej efektywne zapobieganie i wykrywanie
oszustw
Możemy robić to czego poprzednio nie byliśmy w
stanie
Occupational Fraud is estimated to cost
organizations $3.4T annually.
Survey participants estimated that the
typical organization loses 5% of its
revenues to occupational fraud each year.
20% of the cases were greater than $1M.
The frauds reported to us lasted a median
of 18 months before being detected.
2
Oszustwa mają znaczący wpływ na biznes w różnych sektorach
250
Financial
Services
Insurance
200
Billion
150
Retail
100
Healthcare
50
Energy
0
Average Annual Industry Fraud
Losses
Telco
Sources: Banking – BBC News; Energy & Utilities – Electric Light & Power Magazine; Taxation – The Economist; Healthcare – National Health Care Anti-fraud
Association; Insurance – Coalition Against Insurance Fraud; Warranties – Warranty Week; Worker’s Compensation – US Chamber of Commerce; Travel – Business
Travel News
3
Istnieje praktycznie dowolna liczba schematów oszustw
Common
Identity Theft
Improper Payments
Insurance
• P&C Claim Fraud
• Underwriting Fraud
• Disability Fraud
Expense Fraud
Balance Sheet Fraud
Procurement Fraud
• Inventory Theft (Internal)
• Return Fraud
• Gift Card Fraud
Telecom
Banking
• Money Laundering
• Mortgage Fraud
• Check Kiting
Health
• Improper Payments
• Enrollment Fraud
• Unbundling, Upcoding, SNR
Government
• Improper Payments
• Benefit Fraud
• Health Care Fraud
4
Retail
• Cloning
• Toll Fraud
• Subscription Fraud
Utilities
• Supplier Fraud
• Pre-paid Meter Fraud
• Payment Assistance Fraud
Media and Entertainment
• Click Fraud
• Improper payments
• Credit Card Fraud
Life Sciences
• Procurement Fraud
• Employee Fraud
• Medical Bill Fraud
Fraud is a deliberate representation or deception intended to result in financial gain.
Fraud is a criminal act.
Abuse refers to similar actions not proven to be criminal.
Financial Crimes includes Anti-money laundering and cyber-risk primarily for banking
Organized rings
conducting sophisticated
attacks against
corporations for
producing financial gains
Staged Events
Money
Laundering
Improper Billing
Improper
Payments
Providers taking
advantage of public and
private institutions for
the purpose of
improper financial gain
Organized
Opportunistic
Individuals seeking
improper payments by
taking advantage of
private and public
institutions
5
Slip Fall
Arson
Tax Fraud
Medical Fraud
Procurement
Financial
Statement
Expense
Employees creating
fraudulent transactions,
records, and claims to
receive improper
payments from
Employers
Połączenie funkcjonalności adresujących pełen cykl życia przypadku w
ramach procesu zarządzania nadużyciami
Śledztwo
Prewencja
Zarządzanie przebiegiem
sprawy oraz analiza
danych i dokumentacji
Śledztwo
Prewencja
Prowadzenie złożonej
sprawy dotyczącej
nadużycia
Zatrzymanie transakcji
podejrzanej na podstawie
analizy w czasie
rzeczywistym
Zatrzymanie
transakcji podejrzanej
Przestępca
Wykrywanie
Wykrywanie
Analiza zbiorów danych o
transakcjach i
aktywnościach w celu
wyszukiwania nowych
wzorców oraz
udoskonalania
mechanizmów
analitycznych
Poszukiwanie nowych
wzorców nadużyć
Detekcja
Analiza transakcji oraz
podejmowanie decyzji
operacyjnych
IBM Proprietary Intellectual Capital – All Rights Reserved [revision 2013-02-18]
Detekcja
Filtrowanie strumienia
zdarzeń w procesach
biznesowych pod kątem
poszukiwania transakcji
podejrzanych i
przekazanie ich do
dalszej analizy
IBM’s Claims Fraud Solution Framework
Systemy front office
Dodatki branżowe
Modele danych, modele predykcji, reguły, raporty, procesy, dane fraudowe
Real Back
Time Office
/ “In line”
On line
Raporty
Prewencja
Detekcja
Wykrywanie
Śledztwo
Integracja
Analiza
predykcyjna
Selekcja
Zarządzanie
sprawą
Raporty
opercyjne
Wizualizacja
relacji
Dashboardy
Analityka
śledcza
Kanał zwrotny
Alarmy i Akcje
Ewaluacja
Reguły
Podpowiedzi
Reguły
Anomalie
Zarządzanie
decyzją
Identyfikacja
Obszar obserwowanych zdarzeń
IBM’s Counter-Fraud stosuje techniki analityczne na różnych płaszczyznach
Entity Analytics
Predictive Analytics
Forensic Analysis
Content Analytics
8
Retrospective Analysis
Model działania obejmujący całościowo proces biznesowy redukuje
liczbę wyłudzień
Identify fraud
risks at policy
submission
Flag & route
fraudulent
claims at intake
Help Adjuster
ID fraud during
adjudication
Alarmy
Identyfikacja
Identity
Behavior
Exposures
Timing
Locations
Identity
Anomalies
Patterns
Prediction
Identity
Anomalies
Patterns
Text / Content
Social Data
• Number of Claims
within 30 days of
Eff Date from this
Agent
• Claimant, police
officer, and body
shop seen 5 times
previously
Prewencja
• Twitter post
identified and
reported to
Adjuster
Perpetually
analyze loss
data to ID fraud
Wykrywanie
Relationships
Anomalies
Patterns/Clusters
Text / Content
Social Data
• New pattern of
data discovered applied to entire
claims book
Investigate,
prosecute &
recover fraud
Report claim
fraud outcomes
and statistics
Śledztwo
Monitorowanie
Case Management
Unstructured data
Entity Link Analysis
Social Network
Visualization
• Productivity to
handle higher case
load as more fraud is
discovered
Dashboard
Reporting
Geo-Spatial
Trend Analysis
Prediction
• Geo-spatial analysis
of ratio of BI claims
in specific area
IBM Claim Fraud Solution
Zgłoszenie szkody
Silnik analityczny
Analiza powiązań
Model
predykcyjny
Reguły
Tożsamości
Identyfikacja
powiązań
Śledzenie list
Optymalizacja
algorytmów
Konsultant CC
Analiza
tożsamości
Wykrywanie
anomalii
Alarm
Nowe śledztwo
Zarządzanie sprawą
Inteligentne śledztwo
•
•
•
•
Analiza wizualna
Wykrywanie zorganizowanej przestępczości
Współdzielenie wiedzy
Alarmy dla szkód wysokiego ryzyka
Intelligent Fraud
Dashboards
Raportowanie i analizy statystyk
Rozkład jednostek determinuje kto jest kto
•
Kto jest kot? Nie ma znaczenia jak bardzo starają się ukryć
ID 1987839
ID 8987009
Kate Mills-Green
1 Bourne St.
Bolton, MA 01512
Phone: (501)6618044
Silverback
DOB: 12/12/71
Kate Mills
4737 Cimarron Dr.
Bolton, MA 01512
Phone: (978)3656631
DOB: 12/13/71
Identity Resolution
Full Attribution – Entities maintain full attribution
Perpetual Analytics - Real-time perpetually updated,
accurate, high fidelity identities
– Automatic self-correcting
– No data latency
– No refreshes / reloads required
Physical/Digital Attributes
ID 1786616
More Data The Better - Massive Scalability
– Internal & external sources
Katie Green
P.O. Box 12743
– Supports limitless data sources
Clinton, MA 01510
Fuzzy Matching – Seeing Through The Subtle
Phone: (978)3656631
Kinear
1
ID 3335673
Katherine D. Green
4737 Cimarron Dr.
Easton, MA 02334
Phone:(508)278-6019
Zycast Int.
DOB: 11/13/71
Predictive Analytics i Decision Management ocenia roszczenie…
1
Inspektor rozpoczyna obsługę roszczenia ...
1
Główny inspektor ocenia sprawę i rozpoczyna działania analityczne
1
Interaktywne Visual Data Mining jest stosowane do identyfikacji wzorów,
relacji
1
Analiza danych niestrukturalnych odkrywa różne powiązania
1
Analiza śledcza pozwala odkryć nieznane powiązania
1
Główny inspektor analizuje dowody przedstawione przez analityków
1
Streszczenie sprawy jest stworzone do dlaszej analizy
1
IBM posiada mocną historię (strong history) pomagania klientom w adresowaniu
oszustw i przestępstw finansowych
Banking
Insurance
Government
Healthcare
Anti-Money Laundering
Enterprise Fraud
Management
P&C and Healthcare
Claims fraud
Medicare/Medicaid Fraud
Revenue/Tax Fraud
Health Insurance
Claims fraud
IBM has implemented over
70% of the Global Tier 1
Bank’s AML systems
Business Outcomes
2
Reduced fraud by 30%
while improving on AML
Reporting Requirements
40% improvement in
Suspicious Transaction
Reporting
80% productivity saving
Over 60 global Insurers use
IBM software for fraud
IBM has delivered 50+
implementations
IBM i2 has 50+ installations
in Federal Government
Business Outcomes
$17M savings in first 4
months of use
70x faster settlement on
legitimate claims
403% ROI in 3 months
Reduced investigation
referral window from
45–60 days to 1–3 days
IBM has more than 30 client
implementations for
Healthcare payers
Business Outcomes
Identified $75m in fraud
recoupments in the first 12
months of use
One provider charged >800
days worth of billings in a
single year
Stopped more than USD16
billion in fraud in 2012
Business Outcomes
Identified 200 facilities
with questionable outlier
behaviors
Identified >$20M in
potential recoveries
Potential incremental
recoveries of $20M+ per
year
Infinity zoptymalizowało proces roszczeń jednocześnie redukując
oszustwa
403% ROI
95% reduction
in referral of questionable claims
$1 million increase
to the company’s bottom line
Solution Components
• IBM® SPSS® Modeler
• IBM SPSS Collaboration and
Deployment Services
• IBM SPSS Decision Management
Business Challenge: Infinity wanted to submit fraudulent claims to its investigative
unit faster, speeding the payment of legitimate claims and lowering its high monthly costs for
subrogation, the process of collecting damages from the at-fault insurance company.
The Solution: Through automated data analysis and predictive modeling, the insurance
company can more effectively spot suspicious claims early. The solution allows it to expedite
payments of legitimate claims thereby improving customer satisfaction and loyalty while
reducing third-party collection fees.
“I was looking for a product for the enterprise, one that we could use for a variety of predictive
analytics. Primarily, I was interested in speeding the settlement of claims that did not contain
elements of fraud. SPSS was the clear winner in meeting all of our requirements.
- Bill Dibble, SVP Claims Operations, Infinity Property & Casualty Corp.
Segment: IBM Counter Fraud Management
Insurance Bureau of Canada
Reduced efforts
for investigations
Vital information
gathered on fraud
Discovery of fraud
previously unidentified
Solution Components
Business Challenge: Insurance Bureau of Canada recognized that its traditional investigative
tactics were ill-suited to the more insidious and complex fraud threat posed by increasingly
organized networks of conspirators. They knew that it needed to adopt new ways to find and
unravel these syndicates in a faster, more accurate and more cost-effective manner.
• IBM® i2®
• IBM InfoSphere® Identity Insight
• IBM SPSS® Modeler
The Solution: They now have a solution that analyzes not only patterns in the claims data, but
also the hidden social and legal connections that underlie various claimants, accomplices and
even healthcare clinics that are involved in these far-reaching fraud schemes.
“We have demonstrated through this solution how we can improve the effectiveness and
efficiency of investigators by using analytics and visualization technology to make fraud
detection smarter and faster.”
- Rick Dubin, vice president of investigative services, Insurance Bureau of Canada
Firma z sektora finansowego w Europie ma mozliwość analizy i
wizualizacji dużych ilości danych dotycząch połączeń
Analyze
large volumes of data
Strengthen
data security
Highly
intuitive solution
Business Challenge: A financial services company in Europe had no viable way to analyze the
numerous internal and external connections to the company’s servers. Thus, it needed to
implement a powerful analytical platform that could provide insight on the overall security of
its data.
Solution Components
• IBM® i2® Intelligence Analysis Platform
• IBM i2 Analyst’s Notebook®
The Solution: A financial services company in Europe is now able to find and graph a complete
range of internal and external connection data from servers across the enterprise, filter this
information to refine searches, and compare the graphs from previous searches to gain insight.
Using the results of this analysis, the security department can take efforts to strengthen its
data security strategy.
A financial services company in Europe gains the ability to visualize and analyze large volumes
of server connection data and anticipates a stronger data security strategy.
Segment: IBM Counter Fraud Management
A nonstandard insurance company in the United States flags five
times more claims for fraud investigation
10% increase
in the number of claims subject to
special investigation
Expedited
processing of legitimate claims
Expanded
in to new markets by better
assessing regional risk
Solution Components
• IBM® Cognos® 8 Business Intelligence
• IBM SPSS® Modeler Premium
• IBM SPSS Decision Management for
Claims
• IBM SPSS Collaboration and Deployment
Services
• IBM i2® Analyst’s Notebook®
Business Challenge: Adjusters at a nonstandard insurance company in the United States
reviewed hundreds of claims for suspected fraud, yet flagged only 2 percent for further
investigation. The culprit was a manual review process that was not only tedious and timeconsuming, but also failed to identify hidden, more sophisticated instances of fraud. Given the
alarming statistics, the company knew there was a much higher potential of loss it wasn’t
capturing.
The Solution: This insurer is winning the battle with a solution that applies predictive analytics
and a complex suite of business rules against customer and claims data to flag claims
suspected of fraud. As adjusters collect and enter claimant information, the solution
automatically begins rating the claim, applying scores from 1 to 500 based on business rules.
The solution’s intelligence allowed the insurer to redesign the claims-handling process. Now
the company can not only identify claims that most likely require investigation, but can also
give legitimate claims the express treatment.
Narodowa agenacja celne w Europie używa analityki do identyfikacji
schematów oszustw
11% increase
in recovered funds from various
tax evasions
10,000 identified
fraudulent customs claims
Business Challenge: This national customs agency in Europe combed through thousands of
records, from financial transactions to phone records, to detect clues that would help it
identify organized crime rings. This lengthy, manual approach prevented it from uncovering
95% reduction
the subtle or hidden patterns that could lead to identification of suspicious situations. Faced
in time spent collecting data, from a with a growing number and complexity of organized crimes, the agency needed a new
approach to stop rampant tax evasions and smuggling of goods.
week to just hours
Solution Components
• IBM® i2® Analyst’s Notebook®
• IBM i2 iBase
The Smarter Solution: The agency uses a powerful analytics solution to gather,
analyze and link a wide range of intelligence data with no perceivable connection
to uncover patterns of suspicious activities that are part of complex schemes to defraud the
government. Analysts can match and analyze data from disparate
data sources rapidly, discovering connections among numerous aliases, customs
declarations, border crossings and offshore financial institutions.
The agency is now better equipped to stop the proliferation of customs fraud, protecting
citizens and legitimate businesses.
Grupo Bancolombia stosuje uses data mining do identyfikacji
potencjalnie podejrzanych transakcji
40% increase
in identifying suspicious transactions
200% increase
in reporting capabilities
80% increase
in analysis productivity
Solution Components
• IBM® SPSS® Modeler
Business Challenge: To adhere to stricter governmental reporting requirements, Grupo
Bancolombia needed to analyze millions of daily transactions to identify current and
potential fraud.
The Solution: The bank deployed predictive data-modeling software that helped it more
easily and quickly detect transactions that were part of potential money-laundering
operations. By detecting and analyzing expected and typical patterns of over 1.3 million
transactions per day, the solution prevents, detects and reports potentially fraudulent
banking activities that may stem from criminals and terrorists.
“With the data mining system, we generated productivity savings of nearly 80 percent.”
— Francisco Ruiz, Head of Compliance, Bancolombia
Dziękuję!