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
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:
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
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.
Notes de l'éditeur
By the fraud phenomenon. And the pattern is growing outside of Italy too.
Even if it seems that the amount is much more bigger and it’s around 10%.
(none)
These are the most dangerous.
About efficiency on counter fraud management.
Fraud increases costs, the cost increase is overturned on rates, the customer pays.
With particularly sophisticated techniques that are converging even more towards technology.
Notes a study by Rand Institute
False documentations... in order to get discounts not provided
Actions to prepare.. That will be acted during the settlement phase
The company legal and the medical experts involved are not usually considered as a possible threat
The Critical Success Factors determine the performance of counter fraud management
The “objects” are the claims opened daily, as well as the claims already opened that change status during operation.
The technology helps to overcome the problems of the control structure associated with the analysis of massive amounts of data in order to identify a hidden phenomenon and it realizes benefits throughout the entire process of fraud contrast.
Efficiency: with sophisticated and homogeneous procedures in the organizational system
Effectiveness: the technology helps to discover new and sophisticated fraud paths
Each technology provides the control structure of some specific features, complementary to each other, in order to support the different phases of fraud governance process. In our case we have concentrated on the Detection and Investigation phases of the process.
IBM offers a broad and comprehensive portfolio of services and products based on Enterprise Information Management and Business Analytics, conjugated to consulting services organization and process. In order to achieve a concrete counter fraud solution the right chosen approach was the Decision Management.
First off, decision management is a business discipline. It allows organizations to automate, optimize and govern their repeatable business decisions. The people that care most about decision management are business people. Decision management is an extension of the decisions business people would make if they had unlimited time to make those decisions. Technology is the enabler that makes it possible to capture, change and govern decision logic in a controlled and scalable way. Technology also allows these decisions to be automated and called in real-time by processes, applications, and other business solutions.
The key element on this picture is the diamond that represents Decision Services. A Decision Service is an independent, externalized service that can be invoked from anywhere in the enterprise. Supporting Decision Services are two key types of Decision Management, Operational Decision Management and Analytical Decision Management.
Operational Decision Management is based on business rules, and also incorporate business events. This style of Decision Management is generally, but not always, used for high volume decisions, often based on hundreds or thousands of business rules.
Analytical Decision Management is based on predictive analytics, and also incorporates optimization. This style of Decision Management is used to predict the probability of outcomes for which we are not sure but need to make our best educated guess, based on predictive models.
So, Analytical Decision Management is based on predictive analytics and optimization.
Operational Decision Management is about automating repeatable decisions that are subject to frequent changes. Decision automation is an important aspect of decision management but before automating decisions, you have to capture and manage them, at the IT and business level. Allowing business people to take ownership of their business decisions is a key success factor of any decision management methodology.
At any step of the decision making process, organizations are expecting the right decision to be made at the right time. The technology supporting decision automation must provide the required visibility to make business teams confident in the rules that are executing. The technology must provide them with an easy way to access their rules, understand and verify them, and trace their execution.
Decision automation is also about managing the changes to and the complexity of business decisions. Due to economic demands, competitive demands, regulatory changes, or business reasons, rules need changing, requiring flexible tools to modify them in a collaborative team environment. Modifying a decision is usually not a single person’s responsibility. Many people are involved to identify the change, proceed with the change, review it, and validate it - and all within a business compliant timeline. Collaboration is an important requirement that a decision management platform must fulfill.
Business decisions are critical organization assets that must be secured and managed in a controlled environment. Organizations can’t allow just anyone to access and modify decisions that are driving daily activity. Governance is an important challenge that Operational Decision Management platforms must provide, ensuring that decisions are modified by the right persons having the right skills and with the right responsibilities.
These three important challenges are driving IBM in the development of its Operational Decision Management platform: IBM ODM V8.0
This slide shows two examples of how Rules are used in various places of an insurance organization to support decision automation in car insurance application
Problem:
Many applications and processes across the organization are making similar decisions like estimating an insurance price or deciding about a customer eligibility. These decision are invoked from various systems in a synchronous mode but requires to be shared and centralized in order to make the decision maintenance shorter and easier to govern.
The left third of the slide shows the insurance application business process supporting the web site applications. This process invokes 2 decisions: Eligibility and then pricing depending on the eligibility results. In addition the insurance organization has a call center allowing their customers to get insurance quotation on the phone. These 2 different channels are sharing the same pricing component.
Grey arrows: Synchronous invocation of decision services from the process or the call center application
The middle third of the slide shows business rules logic –
The blue text: The business rules that are triggered, based on contextual data transmitted during the service invocation (in grey from the left hand side)
The right third of the slide shows the comprehensive business decision that can be made by the business rules in the middle third of the slide
Let’s have a look on how the IBM Decision Management approach has been conjugated and implemented for Intesa SanPaolo Assicura insurance company, in order to meet their business requirements.
INTESA SANPAOLO ASSICURA S.P.A. IS THE INSURANCE ARM OF THE INTESA SANPAOLO GROUP, WHICH PROVIDES BANKING AND INSURANCE SERVICES THROUGHOUT ITALY.
THE INSURANCE POLE IS ACTUALLY LEADER IN ITALY AND EUROPE, WITH A PORTFOLIO IN ABOUT 68 BILLIONS OF EUROS E OVER 3.7 MILLIONS OF CUSTOMERS.
WE HAVE MORE THAN 4.000 BANK TELLERS IN THE GROUP WHERE TO FIND THE ENTIRE OFFER OF PRODUCTS (LIFE INSURANCE, VEHICLE INSURANCE, HEALTH INSURANCE, CREDIT PROTECTION INSURANCE & PREVIDENCE).
MORE THAN 4200 (FOUR THOUSAND AND TWO HUNDRED) PRIVATE BANKERS PROPOSING OUR INSURANCE AND PREVIDENTIAL PRODUCTS
OVER 300 AGENCIES COMBINE PERSONAL AND BUSINESS CREDITS WITH CREDIT PROTECTION INSURANCE
ONLINE SITE FOR DIRECT SELLING OF OUR INSURANCE PRODUCTS
IN THE LAST FEW YEARS OUR COMPANY DECIDED TO CHANGE SCOPES AND NOT TO BE A TRADITIONAL COMPANY ANYMORE
THE NEW VALUE PROPOSITION IS BASED ON SERVICES FOR FAMILIES, COMBINING TRADITIONAL INSURANCE WITH TEHNOLOGICAL DEVICES ABLE TI GIVE ADDITIONAL SERVICES AND TO FACILITATE COMMUNICATION BETWEEN THE CUSTOMER AND THE PROVIDERS OF SERVICES.
THE NEW OFFER OF PRODUCTS APPROACHES CUSTOMERS IN TWO WAYS:
«BUNDLED» WITH INSURANCE + SERVICES
«UNBLUNDLED» WITH INSURANCE OR SERVICES STAND ALONE
THE SCOPE IS TO MAKE FEEL OUR CUSTOMER THAT WE TAKE CARE OF HIM, TECHNOLOGICAL DEVICES GIVE A SENSE OF SECURITY AND MORE EVEN GIVE A “PHYSYCAL IDENTITY” TO THE SERVICES THEY BUY.
BANK TELLERS AS A KEY ELEMENT TO ACHIEVE GREAT MASS OF POPULATION
OUR DISTRIBUTION CHANNELS ALLOW US TO PROPOSE OUR PRODUTS TO A LOT OF PEOPLE ALL OVER THE COUNTRY
FEW PRODUCTS MODULARS AND FLEXIBLES THROUGH RESEARCH OF TECHNOLOGY AND SERVICES OFFERED BY MARKET AROUND US
THE LATEST GAMMA OF PRODUCTS NAMED «CONME» THAT MEANS «WITH ME», BASED ON A SIMPLE CONCEPT TO GET HELP IN CASE OF NEED.
TO PRESS A BUTTON
ANY DEVICE, FOR CAR O FOR HOME HAS A QUIET BIG RED BUTTON TO PRESS AND GET HELP.
EVEN MORE, IF YOU HAVE AN ACCIDENT WITH YOUR CAR BY CRASHING , THE OPERATIONAL UNIT CALLS YOU IMMEDIATELY TO GET INFORMATION ABOUT YOU AND YOUR CAR AND TO PROVIDE THE NECCESSARY ASSISTANCE
IF YOU DON’T ANSWER AND THE BOX REVEALS A BIG CRASH AN AMBULANCE WILL BE IMMEDIETELY SENT ON THE SITE OF ACCIDENT
AT HOME, IF AN ALARM IS REVEALED AND YOU DON’T STOP IT, YOU HAVE A CALL ON THE DEVICE OR ON YOU PHONE TO GET INFORMATION OF YOUR NEEDING OF HELP.
ADVANCED AND TOTALLY INTEGRATED IT PLATFORM, COMPLETELY WEB-BASED TO IMPROVE SYNERGY AND WORKING QUALITY
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
IN THIS CONTEXT INTESA SAN PAOLO ASSICURA ENGAGED IBM SOFTWARE SERVICES FOR WEBSPHERE AND IBM GLOBAL BUSINESS SERVICES APPLICATION INNOVATION SERVICES TO DESIGN, IMPLEMENT AND DEPLOY A REAL-TIME FRAUD-DETECTION SOLUTION BASED ON THE IBM DECISION MANAGEMENT APPROACH THROUGH WEBSPHERE OPERATIONAL DECISION MANAGER SOFTWARE.
LET’S HAVE A LOOK ON HOW THE BUSINESS NEEDS HAVE BEEN TRANSLATED INTO AN ARCHITECTURAL SOLUTION.
Here you can view the Architectural Overview Diagram which describes the three main layers of the implemented solution, the Fraud Detection & Management System.
When a claim is entered or updated into the company’s claims system, that information, including damage and accident information gathered at accident scenes, is combined with the stored customer data and historical data and finally served up to the Operational Decision Manager software for analysis.
The execution flow consists of two phases: a first synchronous phase of a request submission through Web Service and request placing into a JMS Queue and a second asynchronous phase of the request processing. The second phase starts from the reading of the message from the JMS Queue, continues with the customer, accident and historical data enrichment from customer back-end channels exposed by Web Services and Stored Procedures, proceeds into the Business evaluation logic provided by Operational Decision Management Business Rules and finally ends with the storing of the evaluated claim into customer back-end systems.
The claim evaluation score will be returned to the caller through a callback mechanism and a correlation identifier.
Let’s describe the Front Connection Tier provided by IBM WebSphere Application Server software and the developed Enterprise Web Application.
Let’s describe the Middleware Integration Tier provided by IBM WebSphere Application Server software and the developed EJB J2EE applications.
If the transaction fails, due to back end systems unavailability, it rolls back and it will be repeated for a specific number of times. After that number, the transaction and the claim details will be stored into a JMS Dead Letter Queue for future analysis about the problem occurred.
Let’s describe the Business Logic Rule Tier provided by IBM Operational Decision Manager capabilities and the implemented Business Rules.
Suspicious claims are identified through a set of criteria (business rules) that reflect the expertise of Adjusters and Antifraud Office.
Here you can view some sample implemented business rules grouped by the type of control, such as License Plate Control, Heuristic Control and Person Control, and by the percentage of Activation, that represents the number of current activations of that rule on the total.
This kind of business rules is called Action Rule and the natural language in which they are written derives from a rich combinable vocabulary of words that implement the atomic IT logic. This vocabulary is called Business Object Model.
We have used Decision Tables too, when the pattern of the IF-THEN-ELSE (condition-action) Rule is repeatable and the only key factor of change is represented by the values
The Action Rules, represented by the yellow gear icon and the decision tables, represented by the small table icon are grouped into Rule Tasks.
The Rule Tasks have been orchestrated into a Logical Rule Flow with a starting point and an ending point.
In the Rule Flow you can find Function Tasks that execute some logic like the initialization of some parameters, the flow conditions with different paths to follow and eventually Pre and Post conditions.
Roles like Eligibility, Rule Author, Validator and Rule Deployer have been defined to accurately manage the Business Rules lifecycle
For example if an internal policy changes or if a new Fraud Pattern has been discovered by the analysis of Data Warehouse historical claims
(none)
..Domains like the list of Italian regions and provinces..
Business rules essentially refer to if-then-else statements which include the usage of standard operators. Sometimes the knowledge of a fact cannot be verified to be true or false, for example “the customer is young”. If the customer is 35 years old, then is he young or not?
…..
The bold part of the above phrase could be summarized through the following chart
(none)
(none)
…without the need to know the embedded fuzzy implementation algorithm.
THE PROJECT HAS GONE LIVE ON OCTOBER 2014 (TWENTY FOURTEEN) , LET’S SEE HOW IS IT GOING…. (LEGGI I BULLETS)
THE SYSTEM MANAGES OVER ONE MILLION OF DATA EVERY DAY
So we can discuss about quantitative and qualitative benefits of the implemented Fraud Detection & Management System
Many companies and competitors offer vertical solutions and “ready to use”. IBM approach, project based, has been more innovative and ambitious, less expensive, even if it required more effort by the customer side. The real value of the Fraud Detection Management System is represented by a complete integration of the “well known” Decision Management best practices with a tailored and hi-customizable scoring system, leveraging on the customer Fraud Investigation business context.
A SYSTEM OF DETECTION RULES TOTALLY INTEGRATED WITH CLAIMS APPLICATION AND CASE MANAGEMENT
GREAT SYNERGY AND COLLABORATION BETWEEN IBM AND INTESA SANPAOLO ASSICURA TEAMS, WORKING HARD TOGETHER TO GET TO THE RESULT
BUT MOST OF ALL, ONCE AGAIN, IS PEOPLE TO MAKE THE DIFFERENCE
THANK YOU