The price of fraud by Steven Minsky is a great overview of risk management, especially in how BI can be used as part of an enterprise risk management approach. It is also possible, however, to instantly detect suspicious activity in transactions and thus manage risk in real-time. Taking an enterprise decision management approach by combining predictive analytics, especially neural network models that recognize complex hidden patterns of fraud, and business rules both to capture emerging schemes with no history works well. In addition this approach can minimize unnecessary referrals and interference with legitimate transactions.
One example of this is Falcon One(tm), Fair Isaac's software system for combating existing and emerging types of fraud across the customer lifecycle and at every customer touchpoint.
The Falcon One technology consists of a configurable set of detection, case management, and analytic capabilities to address any fraud type. This unified infrastructure enables multiple fraud solutions to work together consistently, sharing data and system services, to offer maximum protection across every channel and business line.
An EDM approach to real-time risk management, emobodied in software like Falcon One enables you to:
- Respond rapidly to new forms of fraud
- Significantly reduce fraud losses
- Detect and stop fraud earlier
- Reinforce customer trust
- Improve operational efficiencies
- Address fraud within multiple lines of business and channels
So don't just use BI to see what fraud/risk has happened, use EDM to root it out transaction by transaction.