1. © 2015 IBM Corporation
Fraud Detection &
Management System
A real time actionable counter fraud
decision management system
Antonio Dell’Olio – Senior IT Architect
Barbara Camandone – Client IT
Manager

2. Headline
• Frauds in insurance context
• Main fraud types
• Fraud management key factors
• Technology adoption in fraud management
• IBM Decision Management approach
• Intesa SanPaolo Assicura project
• Customer overview
• Customer business needs
• Architectural overview
– Front Connection Tier
– Middle Integration Tier
– Business Logic Rule Tier
• Fuzzy Logic
• Working plan
• Benefits of the solution
• The “X” factor of project success
• Customer Experience
2

3. Frauds in insurance context
3
•The insurance industry is significantly affected, especially on
the motor liability
•Official data amounts to 3% of the total of the fraudulent claims
reported
•More than one third of people hurt in car accidents exaggerate
their injuries
•Studies on the subject demonstrate that over 30% of frauds
comes from the inside
•All insurance companies have anti-fraud measures, no one has an
advanced maturity level
•Fraud feeds a vicious circle
•The Fraud phenomenon is increasing in all regions and is
constantly evolving
•This costs $13-$18 billion to America’s annual insurance bill

4. Main fraud types
4
Underwriting Phase Settlement Phase
Customers
• False declarations
• False documentations
• Actions to prepare the fraud pattern
• Claim request about a never occurred event
• Claim request about an event described with no clear
details
Body Shop
Mechanics
• Invoice reimbursement request about a never
occurred expense
• False billing that does not correspond to effectively
performed services
Insurance
Assessors
• Overestimated assessment or inaccurate damage
reported
• Support to fraud scheme implementations
Counterpart
Legal
• Support to complex fraud scheme implementations

5. Fraud management key factors
5
•Ability to identify the “suspected” objects into massive amounts of
them, in a timely and accurate way
•Ability to deepen timely analysis of information related to the
incident identified as suspicious, leading to the determination of
possible fraud
•Ability to analyze even in false positives, false negatives and
historical claims, in order to constantly improve the detection
methods of undiscovered fraud schemes
•Ability to identify the risks associated with new offers and to
realize the necessary protective measures against fraud, before
launching these offers on the market

6. Technology adoption in fraud management
6
• Efficiency: technology can play rapid and massive processing
• Effectiveness: Investigators analyze only qualified and highlighted
cases
• ROI: significant amounts of resources are held back, compensation is
not paid
• Customer intimacy: the end customer receives equity of treatment
and protection of the relationship
• Dexterity: higher level of service is delivered in a context where the
risk is adequately managed and minimized
Fine Tuning
Governance
DeterrentDeterrent PreventionPrevention Detection Investigation Sanction & Case
Closure
Sanction & Case
ClosureDetectionDetection InvestigationInvestigation
Fine TuningFine Tuning

7. Headline
• Frauds in insurance context
• Main fraud types
• Fraud management key factors
• Technology adoption in fraud management
• IBM Decision Management approach
• Intesa SanPaolo Assicura project
• Customer overview
• Customer business needs
• Architectural overview
– Front Connection Tier
– Middle Integration Tier
– Business Logic Rule Tier
• Fuzzy Logic
• Working plan
• Benefits of the solution
• The “X” factor of project success
• Customer Experience
7

8. IBM Decision Management approach
8
Operational Decision Management Analytical Decision Management
Business Processes, Applications & Solutions
Decision
Services
Business
Rules & Events
Predictive Analytics
& Optimization
Internal & External Data
 Policy
 Regulation
 Best Practices
 Know-how
 Risk
 Clustering
 Segmentation
 Propensity
Scenario Analysis
& Simulation
Scenario Analysis
& Simulation
Decision Management is a business discipline that enables organizations
to automate, optimize and govern repeatable business decisions.

9. IBM Decision Management approach
9
Internal & External Data
How can we ensure that business decisions are managed
in a controlled environment?
Governance
How can we ensure the right decision is being made at the
right time?
Visibility
How can we rapidly respond to evolving market demands,
competitive actions and regulatory requirements?
Collaboration

10. IBM Decision Management approach
10
Internal & External Data
Check customer eligibility
ActActDecideDecide
If driver age is less than18 then
set eligibility to NO
Result = Yes
Customer is Eligible
to the ”First Class
Car Insurance”
InvokeInvoke
Estimate price Insurance Estimated
Price = EUR 350,00
Call Center
Provide 5% discount to gold customers
Insurance application process
Driver &
Car context
Result
If the driver has got more than
3 accidents this year
then flag the driver as high risk
Driver &
Car Context
Result
Invocation of Contextual Decision synchronously from solutions

11. IBM Decision Management approach
11
Internal & External Data
What is a Business Decision ?
Combination of contextual and/or time-based rule artifacts
Contextual Decisions
The externalised decision requires the solution
to provide a well defined informational context
for applying the action rules to.
The decision requires a vocabulary to interpret
the well defined informational context
The result from the action rules is passed back
to the calling solution.
The solution is responsible for taking action
based on the result
Context
Result
Action
Rules
Action
Rules
Action
Rules
R
U
L
E
F
L
O
W
Solution Externalized
Business Decision
Vocabulary
Validation Decision
- Pass/Fail result
ie. Eligibility
Calculation Decision
- Calculated result
ie. Pricing, Tax
Classification Decision
- Multiple results
ie. Gold, Silver, Bronze
Within the externalized decision
1 or more sets of action rules process the
informational context
This is aided by a rule flow to process this in the
appropriate sequence and finalise the result
Different result types require different decision
types

12. Headline
• Frauds in insurance context
• Main fraud types
• Fraud management key factors
• Technology adoption in fraud management
• IBM Decision Management approach
• Intesa SanPaolo Assicura project
• Customer overview
• Customer business needs
• Architectural overview
– Front Connection Tier
– Middle Integration Tier
– Business Logic Rule Tier
• Fuzzy Logic
• Working plan
• Benefits of the solution
• The “X” factor of project success
• Customer Experience
12

13. Customer overview
13
Intesa Sanpaolo Group – Insurance Pole

14. Customer overview
14
Distribution Channels
More than 4.000 «sportelli» of Banks of the Group
More than 4.200 Private Bankers
Over 300 agencies giving Personal and Business
credits
Online site for Direct Selling

15. Customer overview
15
Our Mission
Value proposition based on services for families, enabled by technological
devices
Bank tellers as a key element to achieve great mass of population
New offer of products approaching customers:
«bundled» with Insurance + Services
«unblundled» with Insurance or Services stand alone
Few Products, modulars and flexibles, through research of Technology and
Services offered by market around us
Advanced and totally integrated IT Platform, completely web-based to
improve synergy and working quality

16. Customer overview
16
Innovative Products
Combines traditional Car Insurance with an
assistential component supported by a
technological box with two different purposes:
•Trace the car and intercept accidents
•Direct contact with an operational unit to
provide immediate presence on the accident
site for every necessity
CARINSURANCE
Combines traditional Home Insurance with a
central device and additional accessories for
Safety and Security of your home

17. Customer business needs
17
• As Intesa SanPaolo Assicura grew, our manual, labor-
intensive fraud-review process became increasingly
cumbersome
• The process limited our company ability to investigate
suspicious filings and reduced its overall efficiency in settling
legitimate claims in a timely manner
• We needed to implement a solution that could automate the
fraud-detection process to quickly and accurately spot false
claims
• Any fraud-identification solution had to integrate with the
company claims processing system to automatically stop the
payment process for fraudulent claims

18. Customer business needs
18
•Automate the management of existing anti–fraud indicators
•Identify the suspicions of fraud with a scoring subsystem
•Automate a real-time fraud detection in both phases (underwriting
and settlement) of the claim process
•Allow the customer to free itself from the current service providers
for the management of anti-fraud indicators
•Allow business users to act directly on the operation of the anti-
fraud indicators
•Increase flexibility in the management of the anti-fraud indicators,
ensuring to business users a better and wider information usability,
in order to improve the Time to Market
The solution needs to:

19. Architectural overview
19
J2EE Messaging Bus
Prebuilt Data Model
XSD Schema
Back-end DBMS Repositories
Web ServiceStored Procedure
EJB
Web Service

20. Front Connection Tier
20
•The agency/website/operator caller invokes a SOAP Web Service
exposed by IBM WebSphere Application Server
•A minimum set of key values needed to identify all the claim
details is passed through Web Service input interface
•The web application acquires the request data, transforms it into
XML message and puts into a processing bus JMS Queue
•The application returns to the caller the outcome of the processing
request with a unique correlation identifier needed to the
asynchronous callback mechanism
During the synchronous phase:

21. Middle Integration Tier
21
J2EE Messaging Bus
Prebuilt Data Model
XSD Schema
Back-end DBMS Repositories
Web ServiceStored Procedure
Web Service
EJB

22. Middle Integration Tier
22
•The J2EE Message Driven Bean reads the message from the JMS
Queue
•The back-end Stored Procedures and Web Services are invoked using
the claim identifier data
•The application enriches and combines all returned data about claim
details, customer and historical information into XML data model
described by the XSD Schema
•The application invokes the business rules engine (Rule Execution
Server) using the collected data through EJB Local Interface
•The data are matched against the Business Rules and the customer
Scoring Algorithms
•The returned results which contain the anti-fraud indicators, the
generated score and the request identifier are stored into back-end
systems
•The transaction integrity and coordination is ensured by the IBM
WebSphere Application Server persistence container capabilities
During the asynchronous phase:

23. Business Logic Rule Tier
23
J2EE Messaging Bus
Prebuilt Data Model
XSD Schema
Back-end DBMS Repositories
Web ServiceStored Procedure
Web Service
EJB

24. Business Logic Rule Tier
24
•The collected claim data are matched against the available Business
Rules through the IBM Rete Plus Algorithm
•The Rules are expressed into a natural language that is easy to
understand and to adopt at every level of the enterprise (Business and
IT), so that it will be simpler to enable the change management and the
governance of the Business Rule Lifecycle
•The Rules combine multiple evidences, light on different markers and
compute an overall suspicion score
•A claim is classified as normal, abnormal or suspicious depending on its
score calculated through the integration of the customer sophisticated
algorithm-based scoring system
•The rule scoring system contains the logic needed to detect fraud claims
and it is inquired each time a new claim has been issued, verified or
updated
During the Business Rules evaluation:

25. Business Logic Rule Tier
25
Some sample implemented Action Rules:
License Plate Control
Activation
%
If the input License Plate appears in at least 3 accidents occurred in the last 18 months
then increase the IVASS indicator of 35 points
20%
If the input License Plate appears in at least 1 car accident and its vehicle was classified as
destroyed
then increase the IVASS indicator of 75 points
1%
If the input License Plate appears in at least 1 car accident happened in the last 5 years where the
date of accident is after the effective date of the policy or in the last 15 days of effectiveness of the
guarantee
then increase the IVASS indicator of 60 points
29%
Heuristic Control
Activation
%
If there is no crash report in the list of reports of the black box of the vehicle
then increase the GENERAL indicator of 75 points
63%
If the region of accident is not equal to the region of residence of the insured
then increase the GENERAL indicator of 5 points
5%

26. Business Logic Rule Tier
26
Decision Tables:
Conditions
Each row
is a Rule
Actions

27. Business Logic Rule Tier
27
The structure of a Rule Solution:
Function
Task
Pre/Post
Conditions
Rule
Task
Flow
Conditions

28. Business Logic Rule Tier
28
•IBM Decision Center Portal is
available for Business Users in
order to give them an easy to adopt
rule authoring tool web based
•It is possible to modify rules and
deploy them in real-time (hot
deploy), reflecting the changes
immediately on the counter fraud
patterns
•Integrated security with Intesa
SanPaolo Active Directory and CA
SiteMinder Single Sign On
•Ensures Team Collaboration and
Rule Change awareness, specific
domain vocabulary terms and real
time error detection
Change Management, Rule Governance and Life-Cycle:

29. Fuzzy Logic
29
•We had to deal with Business Rules like:
the customer has been involved in at least 3 car accidents in the last 18 monthsthe customer has been involved in at least 3 car accidents in the last 18 months
•What happens with customer involved in 3 claims in the last 9
months, or customer involved in 3 claims in the last 20 months ?
the customer has been involved frequently in car accidents sometime
around the last 18 months
the customer has been involved frequently in car accidents sometime
around the last 18 months

30. Fuzzy Logic
30
• A “Polygonal Curve of
Approximation” has been
implemented in the context of the
Execution Object Data Model, in
order to approximate other curves
and boundaries of real-life objects
• The underlying idea was to give to
the Business Rules, implemented
into IBM Operational Decision
Manager, the general capability to
use fuzzy quantifiers
• Exposition of new vocabulary
terms (verbalization constructs)
like “around” or “frequently”

31. Working plan
31
• Effort Time: 4 months
• Total Elapsed Time: 7 months
• Periodic progress reports
• Test and Staging environments available before the
Implementation phase
• Production environment available before the Deployment Support
phase
Work Breakdown
Structure
Weeks
Main Phases #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16
Analysis and Design
Implementation
Test
Training & Support
Deployment Support
Training to end users
Business
IT
Milestone

32. What Now ?
32
• Peaks of 3000 Claims in one day
• 19276 evaluated Claims in 4 months
• Data Analysys on 4 different databases (Registry, Policies, Claims,
Blackbox data)
• Over 1 million data per day
• Response time for a complex Claims in less than 3 seconds
• Media of 9000 Claims evaluation a week

33. Headline
• Frauds in insurance context
• Main fraud types
• Fraud management key factors
• Technology adoption in fraud management
• IBM Decision Management approach
• Intesa SanPaolo Assicura project
• Customer presentation
• Customer business needs
• Architectural overview
– Front Connection Tier
– Middle Integration Tier
– Business Logic Rule Tier
• Fuzzy Logic
• Working plan
• Benefits of the solution
• The “X” factor of the project success
• Customer Experience
33

34. Benefits of the solution
34
•The solution, based on IBM products and the expertise of IBM
consultants, has been encapsulated into a repeatable and reusable
asset
•The solution has a flexible, easy to integrate and adaptive
architecture, based on SOA and IBM Decision management
•The solution could be used in other insurance contexts, like Life
and Home insurance
•The solution has extremely increased the fraud detection
capabilities, enabling the development of analytical capabilities
•The solution is modular, scalable, invasiveness and invariant at
any level of the organization
•The solution is maintainable by the existing organization, both
Business and IT departments
Qualitative benefits:

35. Benefits of the solution
35
•Medium claim end-to-end evaluation time: around 9 seconds
•After 6 months of adoption the solution has generated significant cost
savings and real business results
•Anticipates a significant increment of saving up to 60% more annually by
reducing fraudulent claim payments
•Flags erroneous claims in real time, allowing immediate investigation or
legal action
•Identifies emerging fraud patterns so that the insurer can put new rules
and algorithms in place to spot future false claims
•Intesa SanPaolo Assicura has begun to increase its operating margins
as it pays out fewer suspicious claims
•The company has improved its ability to raise customers premiums or
not renew insurance policies in case of repeated fake claims
Quantitative benefits:

36. Headline
• Frauds in insurance context
• Main fraud types
• Fraud management key factors
• Technology adoption in fraud management
• IBM Decision Management approach
• Intesa SanPaolo Assicura project
• Customer business needs
• Architectural overview
– Front Connection Tier
– Middle Integration Tier
– Business Logic Rule Tier
• Fuzzy Logic
• Working plan
• Benefits of the solution
• The “X” factor of the project success
• Customer Experience
36

37. The “X” factor of project success
37
Customer Experience
• Integration
• Synergy
• Collaboration
• People

38. Questions ?

39. Notices and Disclaimers
Copyright © 2015 by International Business Machines Corporation (IBM). No part of this document may be reproduced or
transmitted in any form without written permission from IBM.
U.S. Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with
IBM.
Information in these presentations (including information relating to products that have not yet been announced by IBM) has been
reviewed for accuracy as of the date of initial publication and could include unintentional technical or typographical errors. IBM
shall have no responsibility to update this information. THIS DOCUMENT IS DISTRIBUTED "AS IS" WITHOUT ANY
WARRANTY, EITHER EXPRESS OR IMPLIED. IN NO EVENT SHALL IBM BE LIABLE FOR ANY DAMAGE ARISING FROM
THE USE OF THIS INFORMATION, INCLUDING BUT NOT LIMITED TO, LOSS OF DATA, BUSINESS INTERRUPTION, LOSS
OF PROFIT OR LOSS OF OPPORTUNITY. IBM products and services are warranted according to the terms and conditions of
the agreements under which they are provided.
Any statements regarding IBM's future direction, intent or product plans are subject to change or withdrawal without
notice.
Performance data contained herein was generally obtained in a controlled, isolated environments. Customer examples are
presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual
performance, cost, savings or other results in other operating environments may vary.
References in this document to IBM products, programs, or services does not imply that IBM intends to make such products,
programs or services available in all countries in which IBM operates or does business.
Workshops, sessions and associated materials may have been prepared by independent session speakers, and do not
necessarily reflect the views of IBM. All materials and discussions are provided for informational purposes only, and are neither
intended to, nor shall constitute legal or other guidance or advice to any individual participant or their specific situation.
It is the customer’s responsibility to insure its own compliance with legal requirements and to obtain advice of competent legal
counsel as to the identification and interpretation of any relevant laws and regulatory requirements that may affect the customer’s
business and any actions the customer may need to take to comply with such laws. IBM does not provide legal advice or
represent or warrant that its services or products will ensure that the customer is in compliance with any law.

40. Notices and Disclaimers (con’t)
Information concerning non-IBM products was obtained from the suppliers of those products, their published
announcements or other publicly available sources. IBM has not tested those products in connection with this
publication and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM
products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products.
IBM does not warrant the quality of any third-party products, or the ability of any such third-party products to
interoperate with IBM’s products. IBM EXPRESSLY DISCLAIMS ALL WARRANTIES, EXPRESSED OR IMPLIED,
INCLUDING BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
PARTICULAR PURPOSE.
The provision of the information contained herein is not intended to, and does not, grant any right or license under any
IBM patents, copyrights, trademarks or other intellectual property right.
•IBM, the IBM logo, ibm.com, Bluemix, Blueworks Live, CICS, Clearcase, DOORS®, Enterprise Document
Management System™, Global Business Services ®, Global Technology Services ®, Information on Demand, ILOG,
Maximo®, MQIntegrator®, MQSeries®, Netcool®, OMEGAMON, OpenPower, PureAnalytics™, PureApplication®,
pureCluster™, PureCoverage®, PureData®, PureExperience®, PureFlex®, pureQuery®, pureScale®, PureSystems®,
QRadar®, Rational®, Rhapsody®, SoDA, SPSS, StoredIQ, Tivoli®, Trusteer®, urban{code}®, Watson, WebSphere®,
Worklight®, X-Force® and System z® Z/OS, are trademarks of International Business Machines Corporation,
registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other
companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at:
www.ibm.com/legal/copytrade.shtml.

41. Thank You
Your Feedback is
Important!
Access the InterConnect 2015
Conference CONNECT Attendee
Portal to complete your session
surveys from your smartphone,
laptop or conference kiosk.