January 3, 2012

Well-tuned alert generation and risk models demonstrate good compliance practices to regulators and internal auditors. Regulators and auditors ask how you explain each threshold value in terms of the underlying transactional data characteristics. They check if your selection of alerts reviewed is driven by applicable regulations and policies. They want to know if your decisions to close certain cases, or to continue their investigations, are based on repeatable well defined processes and decision rules.

After the implementation of a compliance or risk management solution such as Actimize, Oracle-Mantas, SAS, Norkom, Search Space or others, two of the most pressing issues are the alerts quality and the huge volume of alerts generated from your monitoring solution. Triage of alerts can take up valuable analytical time and resources – resources you'd rather focus more on actionable alerts and case investigations. Thresholds, filters and scoring are all valuable tools in managing the quality and quantity of alerts. However, they must be carefully applied to ensure you are not "missing anything" and that you can justify them to senior management and external compliance auditors.

You have to make efforts to meet regulatory requirements to tune and optimize periodically your compliance monitoring solution, as well as justify and validate threshold and risk scoring settings. To achieve those objectives tuning and optimization of analytical detection models must be performed by a team professionals with the specialized regulatory data experience.

There are several benefits from periodic tuning of analytical detection models: reduced number of false alerts and missed alerts; compliance analysts operate more efficiently because they spend more of their time on actionable alerts and case development; improved return on technology investments, recommendations for technology and detection model upgrades; expert assessment of system performance relative to current compliance standards; documented guidelines for scenario / model tuning procedures.

To realize expected benefits the model tuning methodology must cover several key area: analysis and visualization of patterns and trends in source and alert data; detection model performance analysis and threshold tuning recommendations; alert and risk scoring rules analysis, recommendations and simulation; analysis and recommendations for enhancements that meet your specific business or compliance requirements; comparison of alert analysis procedures with compliance standards based on regulations, guidelines and best practices.

Exprentis can help you better leverage your investment in your transaction monitoring solutions, as well as enhance the productivity of your compliance analytical team. Read more about our service offerings in this area.