iProtect™ Intelligent Anti-Money Laundering Solution



Money laundering is a diverse and often complex process. Money laundering basically involves three independent steps that can occur simultaneously:

  • Placement - placing, through deposits or other means, unlawful proceeds into the financial system.

  • Layering - separating proceeds of criminal activity from their origin through the use of layers of complex financial transactions.

  • Integration - using additional transactions to create the appearance of legality through the purchase of assets.


In developing iProtect, Brighterion relied on the experience of renowned international money laundering experts. The result are 227 distinct scenarios related to previous known money laundering cases. Similar to a play, a scenario describes and characterizes a money laundering pattern by:
  • a list of players, each one with a specific role to play,
  • a sequence of operations, all of which describe typical money laundering behavior.
The scenarios can be easily updated and improved.

How iProtect Works
  • iProtect "reflexes" trigger the scenarios concerned by any suspicious event,
  • the related banking operations are automatically extracted from the bank's data bases,
  • iProtect confirms as suspicious/unsuspicious the identified scenarios,
  • the system exposes potential money laundering schemes.

iProtect Technologies
The combination of ten advanced artificial intelligence technologies significantly reduce false positives and false negatives. iProtect identifies suspicious transactions with greater accuracy. The ten AI technologies are:

Technology Highlights
Goal-based Multi Agent
  • Allow different technologies to interact and negotiate with each other in order to produce the best result to achieve a particular goal


    		•
    Neural Networks
  • Powerful data-modeling tool that interprets and recognizes complex relationships between data sets
  • Examines historical data and patterns
  • Allows the system to determine if transactions are fraudulent or valid
  • Dramatically improves accuracy

    		•
    Case Based Reasoning
  • Uses knowledge of previous fraud incidence to help anticipate and prevent future fraud opportunities
  • Fraudulent transactions are translated into generic cases
  • New transactions are compared to determine potential fraud incidence

    		•
    Data Mining
  • Ability to address data cleaning, clustering and summarization
  • Searches for relationships and global patterns that exist in a large database
  • Helps the system to detect abnormal transaction behaviour

    • Fuzzy Logic
  • Applies more subtle data interpretation in cases that don't conform to clearly recognized fraud patterns
  • Provides a much lower level of granularity when examining areas that are not as clearly defined
  • Enables users to deploy more powerful expert rules

    		•
    Genetic Algorithms
  • Best suited for solving problems with many variables and a large number of possible outcomes
  • iProtect's genetic algorithm has specific fraud evaluation functions effective in finance, healthcare and telecommunications industries

    • Expert Rules
  • Applies the knowledge of human experts to define conclusions according to the condition presented
  • iProtect uses a patented rules engine that is nine times faster that the industry standard RETE algorithm to access and process business rules

    • Optimization
  • Specific constraints can be added to the system that will define what is allowed or not allowed in a particular outcome
  • iProtect's decision-support (constraints programming) provides tools that allow users to write declarative constraints

    • Text Mining
  • Text mining is the semantic interpretation of textual content
  • Text mining technology is applied to detect money-laundering activity

    • Statistics
  • Provides asymiotic results that can be used to describe likelihood in large samples
  • iProtect uses a number of statistical methods including correlation, regression, Chi-square and others



    • 227 Scenarios and 62 "iProtect Flexible Rules"

    Instead of writing classical rules with limited language and syntax , iProtect is the pioneer in bringing to the fraud and anti money laundering market the power of the fuzzy logic that greatly amplifies the power of human reasoning when dealing with Anti Money Laundering and Fraud prevention.

    iProtect™ Flexible Business Rules is an object-oriented software that allow flexibility, personalization capabilities, rapid application configuration, modification, and deployment.
    iProtect Flexible Rule Builder provide an intuitive browser visual development environment for writing, editing, and testing flexible business rule tool that can be used without any programming skills.
    iProtect provides 62 generic flexible rules that enable financial institutions to monitor in Real time :
    1. Sporadic Account Activity
    2. Large Dollar Reactivation of dormant accounts
    3. No signer relationship
    4. Similar Addresses - Similar signers
    5. Organization goals
    6. Deposit Mix
    7. Use of multiple accounts
    8. Transactions to areas of concern
    9. Who is behind the charity?
    10. Matching inflows and outflows
    11. Transactions with no logical purpose
    12. No logical connection
    13. Demand for greater confidentiality
    14. Foreign Customers
    15. Multitude of account types
    16. Speed of transactions
    17. Volume of transactions
    18. Anonymity of transactions
    19. Compliance efforts of respondent bank(s)
    20. etc.
    iProtect can be used in real time or in a batch processing mode. Benefits
    • Significantly reduce false positives and false negatives.
    • Identifies suspicious transactions with greater accuracy
    • Reduces reliance on manual updating of models and cases
    • Custom models target areas where your business is most vulnerable
    • Provides flexibility to incorporate multiple data sources, add rules dynamically, and deploy multiple 'smart agents'
    • Incremental Learning (Continuously adaptive solution).
    Minimum System Requirements
    Data Server
    For the server machine, hosting the database, the minimum requirements are:
    Operating System : Unix, Solaris, Windows 2000
    Memory: 1GB or greater
    Disk Capacity: 36 GB SCSI hard drive space.

    Client Machine:
    For the client machine, accessing the IProtect Suite remotely:
    • Operating System: Windows 95, 98, 2000, XP, Unix, Solaris
    • Browser: IE5.x and above, Netscape 7.x and above
    • Memory: 64 MB
    • Disk Capacity: not required
    Database Requirements
    The following are the 'natively' supported databases by the iProtect Solution :
    • IBM DB2 UDB 7 and above
    • Microsoft SQL Server 7 and 2000
    • Oracle 8 and above
    Transaction Requirements
    The following are supported transaction environments :
    • TCP
    • HTTPS