Increased fraud detection rate
A combination of expert knowledge and advanced data analysis allows the user to increment considerably the number of detected fraud cases. The use of AI and data mining algorithms makes it possible to identify unusual events, and improve the effectiveness of red flag rules.
Reduction in costs thanks to automation
Automatic data download and risk identification mechanisms reduce the number of manual processes. The system reduces admin to a minimum and accelerates case handling.
A smaller number of false alerts
Based on the history of identified fraud cases, the system "learns" to identify any false positives and fine-tune the search rules.
Optimization of the number of alerts
The use of AI and risk assessments from the case management module allows the user to optimize the system thereby reducing the number of manual risk assessments meaning fewer people engaged in the verification process.
Multistage fraud detection system
Discovering fraud cases in real time:
- Automatic identification of suspicious cases
- Deepened risk assessment implemented within case verification
- Continuous monitoring of the level, areas of occurrence, and detected fraud cases, with the use of dedicated management panels and KPI indicators:
- Monitoring fraud across the sales network
- Monitoring criminal actions performed on the client's accounts
- Monitoring fraud in branches
Tailored solution to meet the requirements of particular sectors
Know-how from PS FRAUD implementation projects undertaken in different sectors allows for accelerated solution implementation, and assurance that customers benefit from the best market practices across many fraud detection areas.
Broad risk factor identification
The system is fed with the institution's own data and the data from external databases, assisting in detecting the risk factors e.g. reserved documents. PS FRAUD automatically processes and updates this data to form black lists, which are central to the fraud identification process.
The data from transaction systems are transferred to the decision-making engine, which evaluates the transactions in terms of the risk of fraud. It automatically alerts the most suspicious cases and sends the case for analysis in the alert management system.
Possibility of multidimensional calibration
PS FRAUD allows for calibration of the scenarios, rules, patterns and models used to identify fraud. This opens up the possibility to utilize ready-made models and adjust alerts to the specified risk scale or to specific business conditions.
Support for the investigation process and documenting fraud cases
For high risk transactions, alerts are automatically established and assigned to specified employees. The system makes it possible to collect and store information and documents behind the alerts, and the entire workflow is automated according to the embedded procedures.
Ready-made rules for many types of fraud
A broad set of ready-made rules, patterns and models allows the user to adapt the fraud detection process to the specific nature of the sector, market segment, product and type of fraud.
Linking rules with prediction models
With prediction models we gain:
- An opportunity for fast identification of new phenomena (anomaly models);
- Increased precision of the rules (profiling, rule induction);
- Advanced rule efficacy evaluation (apriori statistical models);
- A smaller number of empty alerts (optimization)
Correction of prediction models and rules
Models and rules are updated on the basis of information regarding the effect of verification of any observed cases.
Possibility to test and calibrate the scenarios, rules and analytical models
The system is equipped with dedicated interfaces for simulations. This makes it possible for analysts to independently work on the models, without the necessity of programming.
Use of external and internal whitelists and blacklists
The system has embedded automatic information update mechanisms from external databases. It also allows the creation of customized lists.
PS FRAUD allows ad hoc statistics to be generated as well as ready-made reports to be defined, which will be automatically updated and sent out to selected persons in the organization.