Security Intelligence: Should We Send A Guy With A Gun Or A Wrench?

We are kicking off research on security and identity intelligence, which is about understanding risk and detecting abnormal behavior.  One thing is clear: companies don't even *know* what kind of security (SIM, data,  identity, email, etc.) information they should be inspecting to detect security threats and where they should start eating the giant elephant of risk. They clearly need intelligent and automated systems to establish what a normal baseline means in user behaviors and events and then alert on any anomalies - and when they see any changes to normal patterns, understand whether they should send a guy with a gun or a guy with a wrench.  In this research (which will also be the topic of my Security Forum keynote speech) we will look at the interdisciplinary areas between enterprise fraud management, risk based authentication, data protection and identity management. I want to hear about your concerns, issues, and early case studies/solutions in this area.