Real-Time, Cross-Channel Fraud Prevention for

Processors, Issuers, Acquirers and Merchants

As they confront ever-increasing fraud schemes, financial
institutions should consider defending themselves with Index
Smart-Agent technology.
2 The Benefits of Smart-Agent
The economy relies on secure transactions and effective security to maintain a safe and Technology
efficient business environment. The question companies are asking is: “How do we
efficiently eliminate fraud?” 3 Getting a Grasp on Fraud

There is a huge business need for inexpensive, adaptable, real-time, cross-channel 4-5 Powerful Suite of
fraud detection and prevention solutions. However, most existing solutions are built Technologies
on antiquated technologies such as Business Rules, Neural Networks, Data Mining and
statistics. 5 Creating Custom Models

Data Mining technology creates a decision tree from historical data. This method applies 6 Multi-Faceted Protection
the same logic to all entities, although in reality each entity may have a unique activity
pattern. Decision trees are stagnant while fraud schemes continue to evolve. These 7 Point of Compromise
outdated decision trees fail to detect new types of fraud. Detection

Another widely used technology is Business Rules. This technology requires experts to 7 Integration
divide the population considered into categories, and then write rules for identifying
fraud within each category. This approach has two significant limitations. First, there may 8 Everything You Need to
be many different types of individual behavior within each category. Broadly applying Manage Risk
category-specific rules in these cases will result in poor detection and high false positive
rates. Second, rules are “hard-wired” into the system. As a result, they cannot adapt to ever-
changing fraud schemes or data shifts.

Neural Network technology learns from historical data for future data classification. As with
Data Mining, Neural Networks cannot detect any new fraud schemes (that arise daily), and
apply the same logic to all entities even though each entity has a unique activity pattern.
The Benefits of Smart-Agent Technology
Traditional computer programming is purely algorithmic. An algorithm enables the computer
to repeat long suites of logical operations tirelessly and accurately, as long as the algorithm is
correct. As a result, it is unable to take initiative, fix errors or deviate from a fixed line of code.

Consider the following example: while a computer can be programmed to detect viruses, the best anti-virus software still
relies on a team of engineers, a large amount of manual labor and increasing resources. Due to the rigid format of algorithmic
programming, computers are simply not able to learn and adapt to new threats in real-time. The same problem arises in fraud
prevention, anti-money laundering, credit risk prevention, device identification and data breach detection.

Contrary to previous technologies, Smart-Agent technology learns and tracks the specific
behaviors of each entity (cardholder, merchant, device, etc.). As it learns from each entity’s
individual activities, the Smart-Agent updates relevant profiles and makes effective changes at iPrevent evaluates, in
runtime. real-time, behaviors
specific to each
Smart-Agents are the only technology with an ability to identify and stop, in real-time, individual cardholder,
previously unknown fraud schemes. It has the earliest/highest detection and lowest false merchant, ATM, etc.
positive rates because it separately follows and learns the behaviors of each entity.

Rather than being programmed to anticipate every possible scenario or relying on predefined
rules and structured queries, Brighterion’s Smart-Agent technology learns, predicts, infers and
makes intelligent real-time decisions.

2 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Getting a Grasp on Fraud in Real-Time
Detecting and preventing fraud is a dynamic and continuously evolving process. In the area of
fraud prevention, there is a scarcity of effective programming technologies that address the ever
changing security risks and fraud landscape. With Smart-Agent technology, companies can stop
fraudsters in real-time.

Update
Model Incremental
Learning
Algorithms
Input Data

Model
Reconciled
Outputs

Getting a Grasp
Scored on Fraud
Data

Smart-Agent technology is more effective than existing technologies on the market because it allows one-to-one profiling of
entities. This allows it to track every merchant, every card and every device in real-time. The agent follows the unique behavior
of the entities and pulls data from across multiple channels to produce an extremely robust virtual profile that automatically
aggregates all available data and refreshes itself in real-time.

The Smart-Agent uses long-term profiling to establish the baseline behavior of the entity by profiling the behavior over 1 to
24 months. The Smart-Agent also uses real-time profiling to track the same behavior but over a much shorter time period
(a few seconds to a month). By creating these profiles the Smart-Agent automatically compares all behavior against the
entity’s baseline, as well as the baseline for related peer groups. When anomalous behavior is detected, Smart-Agents use this
information to make decisions about fraud in real-time.

3 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Powerful Suite of Technologies
A comprehensive fraud solution must incorporate multiple complementary technologies.
Brighterion’s patented approach combines 10 cutting-edge artificial intelligence technologies.

Smart-Agents Neural Networks

Smart-Agent technology learns the specific behaviors of an Brighterion’s proprietary neural networks can translate any
entity (card, phone, computer, individual, etc.) to create virtual database to neurons without user intervention. Brighterion’s
representation of each individual entity. The virtual agent will neural networks are incremental and adaptive, allowing the
continuously learn from the activities and behaviors of the size of the output classes to change dynamically.
entity it represents to detect any abnormal activity in real-
time.
BRMS (Business Rules
Management System)
Constraint Programming
Brighterion’s iPrevent is embedded with a patented rule-
Constraints can be added to the system that defines what is engine algorithm which is faster than RETE and can accept an
allowed or not allowed in a particular outcome. Brighterion’s unlimited number of rules. iPrevent also provides the tools to
constraint programming provides tools that allow users to allow non-technical experts to write declarative rules.
write declarative constraints.

Real-Time Profiling
Fuzzy Logic
iPrevent uses real-time profiling to evaluate behaviors specific
Traditional and classical logic typically categorizes information to each individual entity (merchants, cardholders, ATMs, etc.)
into binary patterns such as: black/white, yes/no, true/false, Profiling criteria is gathered and updated in real-time for each
or day/night. Fuzzy Logic handles uncertainty in data. Fuzzy entity, giving Brighterion the power to stop fraud before it
Logic is able to handle partial truths in situations where the happens.
answer lies somewhere between what is true and what is false.

Data Mining
Long-Term Profiling
Data Mining extracts knowledge from data. Brighterion’s suite
of proprietary data mining algorithms provides a number iPrevent’s long-term profiling builds a profile for each
of different technical approaches to address data cleaning, individual entity (merchants, cardholders, ATMs, etc.) to
sampling, clustering, learning and classification. Brighterion’s understand their normal behavior. Smart-Agents then use this
data mining can automatically generate business rules from base profile to detect abnormal behavior and prevent fraud.
the customer’s data.

4 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Case-Based Reasoning Adaptive Learning
Case-Based Reasoning is the use of past experiences or Brighterion’s unique, self-adaptive ability allows models to
cases to prevent and detect fraud. Brighterion’s Case-based learn and make effective changes at runtime. This makes our
Reasoning translates transactions to generic cases without solution the only one on the market with the ability to identify
user intervention. The cases created are used to evaluate and stop, in real-time, previously unknown or non-existent
transactions in real-time. fraud schemes.

Creating Custom Models

iPrevent automatically creates a predictive model based on the customer’s historical
information. The iPrevent model is immediately capable of preventing fraudulent transactions,
including previously unidentified types of fraud.

With iPrevent, a model can be created automatically without the need for programming, IT involvement, or knowledge of artificial
intelligence. Brighterion’s development team spent more than 10 years making the design of models transparent and easy to use.
iPrevent works with any relational database to create: The neurons, cases, Smart-Agents, fuzzy sets and data mining parameters
that will best adapt to your business. iPrevent will then generate the fraud prevention engine using: Smart-Agents, decision trees,
neural networks, rules, cases, real-time and long-term profiling and pattern signatures to best protect your business.

The advantages of using these technologies are:

Process Transactions in Best Industry Performance:
Real-Time: Earliest/Highest detection and lowest false positive
Score transactions in real-time (before rates on the market.
authorization or approval) to prevent
losses and protect your brand. Automatic Model Creation:
Automatically create multiple models based
Precise and Specific on region, card type, season, etc. in a single
Behavior Analysis: installation.
Evaluate behaviors specific to individual
cardholders, merchants, ATMs, etc. in real-time. Dynamic Deployment:
Dynamically deploy models into production
Stop Previously Unknown without interruption to the system.
Fraud Schemes:
Unique self-adaptive capabilities detect No Imposed Data Layout:
and stop previously unknown fraud Create and operate models using any
schemes in real-time. dataset for any business need.

5 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Multi-Faceted Protection
A major limitation with current fraud solutions is the lack of real-time, multi-faceted protection.
To efficiently prevent fraud, a solution must provide protection in several areas:

• Endpoint and Navigation:

Uses advanced device ID and
tion Cross
etec
clickstream analytics;
D -Ch
ud an
ne
Fra l F
al sis n
• User and Account:
l y Cross R
na a r i so - C h ea

ra
Employs technologies for real-time
n
ter

u
p a

dP
cross-channel behavioral analysis
Peer ink A
/In

l-T el Fr
m

rev
nn
(specific to each individual cardholder

i m au d
ant

Co

entio
L

e
and merchant outlet);
Merch

n
• Merchant and Outlet:
Provides real-time analysis of the
D e vic

tio n
relationships among internal and
external entities to detect organized or
N et w

te c
h im e

tio n
e ID

collusive criminal activities or misuse.

l-T

De
or k

a
d
an

Re ac

i c
Cl
/D

ic k re

nt
stre B
ev

ce Da t a
de
am I
ch
i

Pr a
o te c re
tio n
Da ta B

In a world of ever-more clever thieves, where new behaviors of fraud arise daily, a smarter, multi-layered fraud prevention
solution is necessary.

With iPrevent customers can manage their entire portfolio (credit, debit, ACH, wire, account takeover, online banking, etc.) on
a single platform. Unlike other fraud solutions, iPrevent integrates a powerful combination of artificial intelligence, machine
learning and Smart-Agent technologies to evaluate, in real-time, behaviors specific to each individual merchant/cardholder
delivering results that are unparalleled in the industry. Additionally, iPrevent’s unique self-adaptive ability allows models to learn
and make effective changes at runtime. As a result, iPrevent is the only product on the market with the ability to identify and
stop, in real-time, previously unknown fraud schemes and offers the longest lifespan on the market. iPrevent processes 6,200
transactions per second in real-time with less than a 10 millisecond response time on a single mid-range server. iPrevent is used
by the leaders of the financial industry worldwide to secure billions of transactions monthly.

6 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Point of Compromise Detection
Smart-Agents also collaborate in order to quickly and efficiently identify phony merchants and points
of compromise.

Specialized Smart-Agents profile cards to keep track of every place they have visited. These specialized Smart-Agents are
activated by high score transactions and initiate collaborative research to find a common point of compromise. For example,
Smart-Agents can be used to detect that a large volume of cards presenting suspicious behavior have all visited the same online
merchant. The Smart-Agents will proactively react, increasing the risk associated with every card that visited that same merchant.

Integration
Brighterion’s platform is based on a distributed architecture. This enables excellent performance using
entry level Unix machines, obviating the need for costly hardware or a mainframe. A distributed
architecture brings several benefits including:

• Lower hardware cost: Billions of transactions per month on a single mid-range server
• Guarantees high operational availability: 99.9999% up-time guaranteed by contract
• High throughput and low response times: 6,200 transactions per second with less than 10 ms response time

This distributed architecture makes Brighterion’s solution iPrevent easily integrates with existing systems and
scalable and resilient to disruption because it has no single infrastructures using industry standard formats such as TCP,
point of failure. ISO8583, Euronet, Web Service, etc. All communication
between iPrevent servers are sent through encrypted TCP
Built-in, intelligent routing allows the users to scale messages, ensuring customer data security.
their system while still making it easy to maintain.
While promoting the benefits of a distributed architecture, iPrevent processes billions of transactions and records per
the iPrevent fraud prevention solution also delivers benefits month from various data feeds and sources. The distributed
from centralized management capabilities. This enables architecture of iPrevent allows us to achieve responses of 10
organizations to monitor the status of any iPrevent server ms on one mid-range server.
from a specified location.

7 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS

Everything You Need to Manage Risk in
a Single Platform
In addition to iPrevent, our real-time, cross-channel fraud prevention solution enables financial institutions to secure their
various portfolios, (CNP, CP, credit, debit, ACH, wire, account takeover, online banking, mobile, etc.). Brighterion also offers a
comprehensive suite of products enabling customers to manage their risk and compliance on a single platform. This removes the
task of integrating solutions from multiple vendors.

Brighterion offers the following products:

iPrevent: Real-Time Cross-Channel Fraud Prevention
iDetect: Real-Time Data Breach Detection Solution
iPredict: Credit Risk and Chargeback Prevention
iComply: KYC and Anti-Money Laundering Solution
iMonitor: Adaptive Learning and Monitoring
iSecure: Real-Time, Behavioral Device ID Solution
iLearn: Unsupervised Learning
iSupervise: Web-based Case Management Solution
iProtect: Cyber/Homeland Security Intelligence

“Brighterion’s state-of-the-art technology sets it apart from competitors. Our products are built using our cognitive computing
platform; the world’s most comprehensive artificial intelligence and machine learning solution that predicts, infers, and makes
intelligent real-time decisions.”
Raymond E. Kendall
3-Time Secretary General of Interpol
Brighterion Executive Chairman

8 - NEXT GENERATION FRAUD PREVENTION SOLUTIONS