e-FraudMart is a fraud database technology, populated with online session traffic and searchable by SQL query as online activity occurs, enabling a revolutionary analytics-based methodology for detecting and preempting fraud in real-time.

e-FraudMart enables financial institutions to identify broader fraud patterns as quickly as the first fraud pattern is identified. This is a dramatic advance over current online fraud analytics, which require scrubbing and extracting of historical session logs into expensive data warehouses for after-the-fact reporting. e-FraudMart not only provides rapid intelligence to preempt further fraud, but also enables institutions to block related transactions that may have already have been completed online, but have not yet cleared the bank.

e-FraudMart is an extension of the Business Signatures Real-Time e-Fraud Detector (RTFD) product for in-session behavioral fraud identification and intervention in an online session. e-FraudMart is currently in use at leading US financial institutions that have selected the RTFD product.

e-FraudMart is automatically populated by the Business Signatures data streams Intent Processor. Analytical capabilities include the following:

• Real-time analysis of all online access patterns, including information based on geo-location
• Real-time analysis of all online behavioral patterns, including "in-session" account hijackings, bot attacks, and multi-session account takeovers
• Customizable pre-defined fraud reports, as well as ad hoc query capability using any SQL reporting tool
• Selective sharing of fraudulent patterns in real time to a shared fraud database

According to George Tubin, senior analyst, TowerGroup, "Preemptive defense against online fraud is the new frontier for protection of banking and other financial accounts on the Internet. The sooner a financial services provider recognizes a fraud is occurring, the more effectively it can block intrusions and stop transactions, even those already completed online but not yet cleared. This type of just-in-time knowledge requires real-time investigative ability across transactions and accounts."

Stream Processing Technology for e-Fraud Prevention

Business Signatures e-FraudMart extends the capability of the Business Signatures data streams Intent Processor, the engine developed by Business Signatures in collaboration with universities including Stanford and UC Berkeley. Unlike typical data management systems which provide store-and-query functionality, stream processing operates on a "query first and store later" basis to analyze high volumes of incoming streaming data.

As streams are processed, data is normalized very efficiently in compressed forms, called business signatures. These business signatures represent activity types such as "login", "transfer money", "view check image", including transactional and behavioral activities. The business signatures can then be analyzed on a continuous, real-time basis via standard queries or ad hoc searches.

"As in any risk management effort, fighting fraud often comes down to grasping the knowledge you already have in time to use it. The addition of e-FraudMart to our flagship RTFD product opens up the capability of our platform in a truly revolutionary way," said Peter Relan, chairman and CEO of Business Signatures. "From the moment a questionable pattern is recognized, our e-FraudMart is capable of seeking matches in the stored data to identify organized fraud schemes. The breadth of research capability, as well as the potential to share this information with other financial institutions in real time, is a breakthrough in fraud defense."