COMPARATIVE STUDY OF SAS FRAUD MANAGEMENT & MONITOR PLUS ACF
Agenda
Introduction SAS Fraud Management SAS Functionality Fraud Framework Technical aspects of SAS Monitor Plus ACF ACF Functionality Fraud framework ACF Technical Aspects Comparison of SAS and Monitor Plus
Introduction
Banks need to have enterprise-wide fraud management system
Fraud detection, Alert, Case management should be managed as a whole
Customer behavior needs to be studied across all channels and products
Equal emphasis should be given on non-financial transaction
SAS Fraud Management
Proactive approach Focuses more on fraud prevention than fraud
detection Wide coverage of transactions
Purchase Payments Non-monetary
Monitors real time transactions Multiple accounts monitoring for same
cardholder
Functionality
Integrated End to End Fraud management approach
Functionality
Analyzes customer behavior Scores transactions on demand for risks In built fraud models available System updates itself after every
fraudulent transaction Separate Case Management provided
Fraud Framework
Data analysis and alert generation The ability to assimilate data from multiple sources and
apply predictive analytics to accurately assess transactions, activities and customer state in real time.
Alert management The mechanism for accepting, prioritizing and distributing
alerts from the various fraud detection and money laundering tools used across the enterprise.
Social network analysis An analysis and visualization tool for uncovering
previously unknown relationships among accounts or entities.
Case management A structured environment in which to manage
investigation workflows, attach documentation and record exposure and losses, while using advanced dashboards for management oversight and analytical reporting to track financial crimes operational performance.
Technical Aspects
Zero foot print technology End users need not install any software
Use of neural network Ability to learn Ability to generalize
Client configurable API Can be customized as per client’s need
Customer state vectors Flexible Architecture
Monitor Plus ACF
Reactive Approach Transaction Screening, monitoring
Covers real time, near real time, batch mode
Scope of transactions Cash Advance, Purchases Internet & electronic channels Non financial transactions
Users can customize pre-define parameters
Functionality
• AI Techniques• Unusual TRx• Fraud
Sequences
• Generate Alerts
• Pre defined actions
• Investigation Work Flows
• Confirm• Discard• Fraud
Learning Analyze
ActionConclusion
Functionality
4 stage analysis process ACF analyzes operations on following criteria
Issuer Acquirer Processor Switch
Comprehensive Fraud pattern analysis Automated alerts and actions Uses AI Techniques for analysis
Business rules Case base reasoning Risk factor analysis Transaction risk scoring
Framework
Alert & Presentation Layer
Reports
Analysis Layer
Risk Factors
Data Access Layer
Real time, Near real time, Batch
Framework
Data Access layer Access data from transaction from any kind of source
i.e. real time, near real time, batch mode and screens transactions for fraudulent processes
Analysis Layer Performs analysis based on adaptive business rules,
neural n/w, risk factor scoring, fraud patterns Alert & Presentation layer
Generates alerts such as email, SMS, call to customer service as well as customer himself, Audit etc.
Performs pre defined actions given by users such as block card, deny access, freeze an account etc.
Technical Aspect
Use of neural network Ability to learn Ability to generalize
Customizable system parameters Integrative and flexible system Included case management Point of compromise management
Face to FaceParameters SAS Fraud Management Monitor Plus ACF
Approach Proactive Reactive
Transaction Coverage Real Time Real Time, Near Real Time, Batch Mode
Functionality
Customer Behavior Present uses customer state vectors
Present uses fraud pattern analysis
Transaction Scoring Present Present based on risk factors
Built in fraud Models Present Present
Case Management Need to buy separately Present
Technical Aspects
Zero foot print technique Present Absent
Neural network Present Present
Customizable API Present Present
Customer state vectors Present Absent
AI Techniques Absent Present
Alert management Present Present