A new family of products and services from Unisys has been announced which should make it possible for financial institutions to detect more complex kinds of fraud and spot identity theft and money laundering schemes earlier.

The company unveiled the Unisys Active Risk Monitoring System (ARMS), a suite of software that it claims improves the ability of banks and other organizations to spot the patterns of seemingly unrelated events that add up to criminal activity.

Unisys would rely on technology from Actimize of New York to provide the underlying analytics technology for the new system.

Actimize has developed technology that enables organizations to do complex data mining and analysis on stored information and transaction data without needing to copy data to a separate data warehouse, according to Eli Katz, practice director in Unisys' Global Financial Services group.

Actimize's technology can monitor an enterprise's transactions in real time, identifying elements that may affect models of suspicious behaviour while flagging transactions according to predefined criteria, according to Actimize.

For each transaction analyzed, the system was capable of tweaking its underlying statistics and checking those transactions against the rules used to identify suspicious activity. After being analysed, the transaction was discarded, Katz said.

The Actimize software ran on the Microsoft Corp. Windows NT or Windows 2000 platform and can be deployed on standard server hardware with either four or eight processors, Katz said.

The Actimize technology makes ARMS a more powerful and flexible risk monitoring system than those that rely on large data warehouses, according to Katz.

In contrast to those systems, Unisys ARMS tied together live transaction activity with information stored in a wide range of separate enterprise data stores such as CRM (customer relationship management) systems, enabling organizations to spot patterns of suspicious activities within 1/1000 of a second of their occurrence.

As an example, Katz described a scenario in which a criminal approaches an ATM (automated teller machine) kiosk with 30 stolen or fraudulent ATM cards, using each in succession to withdraw US$500.

Although none of those transactions, taken alone, would be adequate to raise a flag, Unisys ARMS was able to detect a change in the rate of transactions within a given window of time, or spot the increased number of cards that have never been used at that ATM before.

Alternatively, the system could correlate the activities of a number of money launderers who make small deposits into different accounts only to transfer money into one or more consolidated accounts, flagging the entire group as a threat, Katz said.

Suspicious events could trigger either a visual notification sent to an administrator's screen or an action such as shutting down an ATM or group of ATMs, he said.

On top of Actimize's technology, Unisys was contributing its expertise in the financial services industry, developing the logic and underlying statistical profiles that are the guts of the new system, Katz said.

The system typically costs about $1 million to implement, including the price of the Actimize software and services to develop statistical models to fit the customer's needs, Katz said.