Economic losses from issuers, caused by fraud in Credit Unions, have grown despite the implementation of chip cards.
Fraud changes and adapts, maintaining its impact in the financial sector, with issues with “card not present” transactions being the biggest trend.
The analysis of both unusual and normal behavior of cardholders, coupled with the discovery of unknown topologies of fraud through models based on machine learning, are now more relevant than ever.
Sentinel supports the entity's fraud prevention strategy through a real-time approach that allows the authorization process of the transaction to be interfered with other risk criteria, thus becoming a very important filter to avoid losses due to fraud.
It allows the implementation of the strategy of fraud prevention through an analysis from multiple perspectives by different technological components.
Generates predictive models using Machine Learning, assigns an expert score to the transaction and enables the construction of business rules.
Improves the efficiency and performance of your research team, as it allows you to define the workflow and actions to be taken during the analysis and review of a suspicious account.
Origination fraud is an important event that affects the issuers. Sentinel supports the issuing entities in the process of linking cardholders.
Restrictive Lists: Functionality that allows the entity to access multiple restrictive lists or international sanctions lists.
Link Analysis: Enables the discovery of hidden patterns and relationships graphically, facilitating research processes, as well as the detection of organized crime networks. Additionally, it allows detecting fraud in the different stages of the business process, such as sales, origination and delivery.
The Credit Union must have the necessary tools to be able to reject a transaction that has enough characteristics to be cataloged as fraud at the moment of the authorization. This is the first barrier to fraud.
Real Time: Ability to evaluate complex events through a transactional analysis in real time that allows the evaluation of the transaction based on different conditions such as parameters, statistics and lists.
Sentinel gives an answer to the authorizer to continue or stop the transaction.
Cardholders perform a set of transactions, some are approved while others are denied by the authorization system. Sentinel analyzes all the transactions using different technologies to identify unusual behaviors and fraud patterns:
Predictive Models based on “Machine Learning”: Visual design environment to quickly build predictive analytics models for fraud prevention. Provides a complete library of learning algorithms, data preparation and exploration, validation tools and model evaluation.
Models based on Rules: The system presents a modern fully graphical rules editor that facilitates the construction of models for the user: groups conditions, includes lists, multiple statistical conditions, and complex analysis patterns, among others. Expert users can also use formulas on conditions, which facilitates the manipulation of data and results of the rule.
Statistical Behavior: It allows for the generation on individual profiles for each alert objective (associated, cardholder, BIN, etc.), defined as statistical indicators based on the transactional behavior of each of them. These profiles can be used as input when configuring fraud detection models based on rules.
Management: The system permits a series of preventive actions online:
1. Preventive card blocks
2. Blocking ATM withdrawals
3. Decrease available funds in the account
Expert Risk Score: Allows the establishment in the system of a risk matrix to assign a score to each transaction, based on the information contained in the transaction’s plot, for example the MCC, the amount, the time, the entry point, response code, the country of origin, etc. Based on these criteria, a score is assigned according to the values with which the transaction complies.
Geolocation: Analyzes different aspects related to the usual location of customers, service providers, types of connection, among others. IP access data used by users and registered in Sentinel can be used in forensic or criminal investigations, in audits and as evidence in criminal proceedings.
Analysts have a query or viewer that allows the visualization of unusual activity with multiple inputs for decision making:
a. Filter by date range
b. Alert objective
c. Filter per stage of alert
d. Advanced filter
e. Transaction evaluation scores
f. Behavior profile
g. Transaction alert models
Supervisors and analysts may have strategic control of the entire operation of the fraud prevention department, observing the major indicators:
a. Pending alerts
b. Average attention time
c. Alerts marked as fraud
d. Alerts raised to investigation
e. Alert resolution schedules
f. Fulfillment of the goal
Once the analyst reviews the activity that the system suggests is suspicious, you can generate cases that allow you to track suspicious transactions and record all actions during the investigation process.
The case management facility allows traceability to the entire process of managing the unusual activity of the cardholder:
Provides a control panel that allows visualization of:
Powerful tool for data discovery and information analysis. Its main characteristic is that it is aimed at end users without significant technical knowledge, who can design their own reports and dashboards, schedule their execution and send them automatically to different recipients if desired:
The solution includes tools and strategies based on technologies known as “Business Intelligence” and “Data Warehouse” that integrates all corporate fraud information, allowing you to manage and analyze every aspect of your business and your environment, from the beginning to the end.