TFRS-C Transactional Fraud Reduction System for Cards.

TFRS-CTFRS-C

TransactionTransactional Fraud al Fraud

Reduction Reduction System for System for

CardsCards

TFRS-CTFRS-C



CardsCards

TFRS-C

AgendaAgenda

The Company

Product Description

TFRS-C Market Drivers

Functional Overview

Demo

Conclusion

TFRS-CTFRS-C

The Company Product

Description TFRS Market

Drivers Functional

Overview Conclusion

TFRS-C

Sigmenta’s Operating ModelSigmenta’s Operating Model

Headquarters in Madrid, Spain

Majority shareholder is GMAST Group of Israel

Dedicated to the research, development and

implementation of technologies for synchronous

processing and evaluation of transactions (STT ®)

Operating in:

Europe

North and South America

TFRS-C

Spain

Software Development and Sales

Offices in Europe, USA and Latin America

Israel

Presence and projects in 27 countries

Physical Security

Border/Perimeter Security

Turnkey Security Projects

Israel

Group R&D

Venture Capital

GMAST InternationalGMAST International

TFRS-C

Venezuela

Integrated Security Services

Product Representation Services

Security Operations Outsourcing

Research and Intelligence

Venezuela

Software Factory

In-house development and deployment of GMAST applications.

Off-shore outsourced development services for clients

GMAST GroupGMAST Group

TFRS-C

TFRS

Transactional Fraud

Reduction System

Integrated real time solution for

card fraud control

CPS

Card Personalization

System

Personalization of the consumer profile

of payment cards

eBONUS

CNB

Banking services for unbanked

individuals, payment card

personalization, and streamlining of

expense management processes

LoyaltyControls the level of customer loyalty in

financial and non-financial institutions

Our ProductsOur Products

TFRS-CTFRS-C

The Company Product


Drivers Functional Overview

Conclusion

TFRS-C

TFRS Product FamilyTFRS Product Family

This product family is an innovative and robust

solution for protecting card issuing institutions and

their customers against unauthorized use of

payment cards

Uses sophisticated self learned and continually

evolving neural-net rules

Real-time transaction verification

Applies customer behavior and emerging fraud

patterns

TFRS-C

STT 4G STT 4G

Synchronous Transactional Technology 4th Generation -

STT 4G ®

Real-time technology for solving real life problems

No time-out in response

Breakthrough solution for managing transactions

Flexibility to create innovative business models

Strong customer acceptance

TFRS-C

Supports financial institutions’ strategies

Cost reduction

New service development

Customer fidelity

Expands business opportunities beyond the financial

world

Welfare

Government benefits

Telecom

“We created an emerging technology…..”

STT 4G STT 4G

TFRS-C

S.A.T. divides the work into two separate processes:

ASYNCHRONOUS MODULEAnalyzes and processes

asynchronously, without impacting critical response time, all data to be considered in the

client’s next transaction.

SYNCHRONOUS MODULEEvaluates the incoming

transaction in real-time, subjecting it to constraints set by

the asynchronous module.

To evaluate an operation in real-time, we face:

Long code High intelligence Much processing time

Short code

Little intelligence Little processing time

To solve this problem and obtain good results online we need:

High intelligence Little processing time

Short code

Neural Network

&Expert

SystemsTechnology

Synchronous/Asynchronous Synchronous/Asynchronous Technology – S.A.T.Technology – S.A.T.

S.A.T.

TFRS-C

What is TFRS-C?What is TFRS-C?

TFRS is a 4th generation antifraud system that provides real-time analysis of a client’s behavior patterns combined with emerging fraud tendencies.

Based on neural nets and expert systems

Combines over 2,500 algorithms

Takes into account the client’s transactional history over the last 24 months

Analyzes actual transaction within 5 milliseconds (on average)

Flexible enough to accommodate unusual but legitimate transactions

TFRS-C

Continuous self-learning of client’s behavior in

addition to recent fraud patterns

Complex calculations performed in non-critical

time, enabling rapid transaction analysis in real-

time

Applicable to all transactional media, allowing for

rapid turnover

Provides recommendations for executing manual or

automatic operations

What is TFRS-C?What is TFRS-C?

TFRS-CTFRS-C

The Company Product


Drivers Functional

Overview Conclusion

TFRS-C

TFRS-CTFRS-C MarketMarket DriversDrivers

Processors: Many card processors do not possess the technology to provide sufficient protection from fraud

Banks: Reticence - slow to bring cases to the attention of the

regulators, possibly due to reluctance to expose vulnerability Skill Gap - Banks vary widely in fraud detection expertise

Associations: Visa International cannot access or control Merchant Account

information. Only Banks have that information. One wonders what a crooked bank could do with Merchant Accounts.

The distributed nature of the Visa/MC system, with each bank managing its own "business", is a potential weakness

Tools: Need for an intelligent tool to prevent losses, secure electronic monitoring online

TFRS-C

Type

Losses in 12

months to 30

June 2002

Losses in 12

months to 30

June 2003

Losses in 12

months to

30June 2004

Counterfeit £161.1m £125.9m £123.0m

Card-not-

present£108.0m £107.8m £138.8m

Lost/stolen £114.4m £107.1m £118.8m

Mail non-

receipt£32.5m £40.5m £61.2m

Identity theft £15.4m £24.6m £37.0m

Total £431.3m £405.9m £478.8m

Within this

total: cash

machine fraud

£24.8m £33.0m £61.1m

Fraud losses on UK-issued cards

1994£3.2m

1995 £3.5m

1996 £4.4m

1997 £8.2m

1998 £9.7m

1999 £12.2m

2000 £18.3m

2001 £21.2m

2002 £29.1m

2003 £39.0m

Total (1994-

2003)£148.8m

APACSAssociation for Payment Clearing Services

Fraud is a Major Concern for Card Fraud is a Major Concern for Card IssuersIssuers

TFRS-C

Mistakes in Risk ManagementMistakes in Risk Management

NextCard: Application fraud created excessive losses Lesson learned – Verify applicants

Providian: Targeted marketing to customers with poor credit during “good” times created a disaster during “bad” times Lesson learned – Consumers are impacted by economic

conditions

Advanta: Targeted marketing to balance transfer customers in need of credit ballooned into a sea of red ink Lesson learned – Spread portfolio risk

Allfirst: The lack of good internal controls and a lax commitment to internal auditing resulted in a $690 million loss Lesson learned – Spend money on proper controls and

auditing

Most are out of businessMost are out of business

TFRS-C

Conceptual Framework for Risk Issues Conceptual Framework for Risk Issues in Credit Cardsin Credit Cards

Fraud

Application Fraud

Granting of credit Prime and sub-prime

Profit

Origination Portfolio

Adapted from Tower Group, April 2001

Authorization Fraud

Purchase

Origination Risk

List procurement Credit scoring Counter offers

Credit Risk Attrition Delinquency/

Bankruptcy Charge-off Line-of-credit reductions Proactive intervention Lifetime customer value

Sigmenta’s present focus

Sigmenta’s present focus

Credit card risk can be thought of in two components: Credit card risk can be thought of in two components: origination and portfolio – and each has a fraud and profit origination and portfolio – and each has a fraud and profit

risk componentrisk component

TFRS-C

d

Traditional Systems

TEC

HN

ICA

LLO

GIC

AL

Increase in security• in cards• in communications• in user identification

Establish profiles based on statistics

Reactive deferred correction

3rd Generation

SystemsBased mainly on static profile of “fraud type” or “user profile”

Based on if-then rules

Proactive detection

Asynchronous preparation of fraud indicators

Real time detection

4th Generation

Systems

Decision Support System EvolutionDecision Support System Evolution

TFRS-C

od n

Problems with 3Problems with 3rdrd Generation Generation SystemsSystems

Evolved 3rd Generation Systems:

Real-time analysis of previous operations combined with details of the current transaction.

Technical resolution problematic:

- Must be solved in milliseconds- Complex analysis takes too long

to be performed in real-time

Real-time detection, combining client’s history and current transaction

TFRS-CTFRS-C

The Company Product


Drivers Functional

Overview Conclusion

TFRS-C

Synchronous Evaluation

Behavior Profiles

EX

PER

T S

YS

TEM

S

d

N

EU

RA

L

N

ETS

TFRS

Fraud Aspect

Combines more than 2500 algorithms

Includes 24 months of customer transaction history

Synchronous transaction verification (5-10 ms on average)

Asynchronous data analysis (in non-critical time)

44thth Generation Anti-fraud Generation Anti-fraud SolutionsSolutions

External Feedback and Data Mining

Fraud Profiles

Customer Aspect

Continuous acquisition and learning

On-line transaction evaluation

TFRS-C

TFRS-CTFRS-C Architecture Architecture

Intelligent Interface to External Systems

Synchronous Decision Makers

Internal & ExternalFeedback

TFRS front-end application Credit-Debit Card System

orCore Banking Application

Transaction

Authorization System

TxSynchronus

Routines Factors

Asynchronus

Tx

Synchronus Connector• Data Formatter• Basic Logic decision • Absence response• Transactions to discard• Switch’s Evaluation• Message's Services

Asynchronus Connector• Data Formatter• Authorization Data• Combined information

• General Conditions Evaluation• Specific conditions evaluation

Define RoutineFactors Evaluation

• Lock the information• Return Response

ProductOperationChannel

Time

• Specific conditions – Reconfirm Synchronus Evaluation (Regenerate Routine y Factors)

• Postulate Alert• Prepare the next transaction• Demon postulated

Demons• Generate Virtual Transaction• Prepare the next transaction

Front Feedback• Re-calculate Short Term

Coefficient (immediate)

Band Feedback• Re-calculate Mediun Term

Coefficient (1m)

Tx

Data Base

Front-End

Call Center Tecnhical Support Office Branch V.I.P.

Historic Feedback• Re-calculate Long Term

Coefficient (4m)

TFRS DB contains profile indicator

Virtual Operator

TFRS-C

TFRS SYSTEM SCHEME

Trigger

TFRS Database

Synchronous

Alert Engine

LinkAsynchronou

s

ExternalFeedbackInterchange

Clients

Connectors

Platforms

AUTHORIZATIONSYSTEM

Authorization

Tx

CHANNELS

POS

ATM

VRU

CallCenter

Teller

Franchises(Visa, MC,

AMEX)

Internet

Front Operator

s

PersonalizationClient

Managment

Back

Virtual Operator

Console

Front Clients

Inbound/Outbound

Web Pages

Mobile Telephone

Start

Synchronous

OperativeTables

AdministrativeTables

Asynchronous

Card Admin.

Data mining feedback

Single Entity IntegrationSingle Entity Integration

TFRS-C

CHANNELS

POS

ATM

VRU CallCenter

Teller Internet

Franchises(Visa, MC,

AMEX)

. . . . .

Front System

AsynchronousBack

SystemWeb Link

DB TFRSTFRS Centralized System Scheme


Synchronous

Table

Host Bank 2

Synchronous

Asynchronous

Connector

Host Bank 1

Synchronous

SynchronousTable


AsynchronousConnector

Data MiningFeedback

Multi-Entity IntegrationMulti-Entity Integration

Synchronous

Table

Host Bank n

Synchronous

Asynchronous

Connector


TFRS-C

AMOUNT

Establishes allowed transaction amounts, defined by the customers behavior or by the institution

CUSTOMER

Establishes general and/or particular parameters, that define the profile and behavior of a client.

MERCHANT

Sets suspected fraud merchant and also suspected or high risk categories

TIME

Establishes the allowed time for a transaction, and intervals between transactions performed by a single customer

LOCATION

Establishes the expected location for the next transaction, keeping track of previous use or notification

TYPE OF TRANSACTION

Sets the conditions on the type of transactions expected from the customer: cash advance, purchases, etc.

CHANNELEstablishes general and/or particular parameters to define channel-specific transaction evaluation

TFRS-C

TFRS-C Data ConceptsTFRS-C Data Concepts

TFRS-C

end ofmonthWorld 1 World 2

World 3

World 4

A World is defined as the set of rules, algorithms and live transactions of a customer within a certain time frame. Through these rules, we obtain control values that correspond to behavior profiles / time frames.

World SchemeWorld Scheme

TFRS-C

Front and BackFront and Back

Integrates easily with existing systems.

Restricts simultaneous access

The module performs these operations:

Registers fraud confirmation/non-

confirmation

Catalogs fraudulent/non-fraudulent

transactions

Feeds information back to the statistic

modules and neural net.

Transfers actions to the Payment Means

Back Office

The Front & Back module allows the institution to access TFRS through a friendly Web interface. This module:

TFRS-C

The F&B module allows the institution’s operators to access TFRS through a friendly Web interface

Pro

du

ced a

lert CONFIRM

Obtain the necessary information to contact the cardholder

REGISTER

Obtain customer feedback in the system Feedback TFRS ID +

ACT

Select appropriate action, informing the back-officefraud analyst

Respond automatically to fraud alerts without contacting the customer, based on the user’s profile and

on transaction verification history.

a) Through Operators

b) Through Virtual Operators

Front and Back: Work FlowFront and Back: Work Flow

TFRS-C

TFRS

internalfeedback

externalfeedback

• Complements the information by gathering data from external systems. • Generates knowledge about the transactional behavior of each client.• Adjusts and refines fraud tendencies.• Minimizes deviations in the Model’s predictions.

Learning data from outside the system’s processes, such as: Data on card distribution Date of issue Distribution flow, stages Data on fraud history Data from franchises about cards

Learning from results created within the system.

Feedback SystemFeedback System

TFRS-CTFRS-C

The Company Product


Drivers Functional

Overview Conclusion

TFRS-C

TFRS Efficiency IndexTFRS Efficiency Index

Regardless of asynchronous mode, the Virtual Operator re-evaluates alerts periodically

HUMAN OPERATORS

VIRTUAL OPERATOR

INTERNATIONA

L

STANDARD

TFRS

FALSE POSITIVE

Success rate per total alert1 of each 25

1 of each

14

FRAUDULENT

TRANSACTIONS

Average fraudulent

transactions per affected

account

3 to 5 2 to 3

PROTECTED BALANCE

Percent of balance

protected in a fraud

impacted account

55% 82%

TFRS-C

ConclusionConclusion

With TFRS-C, financial institutions can significantly

reduce the rate of payment card fraud, enabling:Cost ReductionImage improvement – maintain a good

reputationResults to date exceed franchise expectations

False positive: 78% less than with previous fraud identification system

Identification of actual fraudulent transactions: 150% more than with previous system

Loss of funds due to fraud: 47% less than with previous system

TFRS-CTFRS-C



CardsCards

TFRS-CTFRS-C



CardsCards

TFRS-C Transactional Fraud Reduction System for Cards.

Documents

tfrsc stt

tfrsc tfrs product family

tfrsc sigmentas

tfrsc agenda

g slide

clients gmast group

south america slide

synchronous processing