Pivotal can look deeply into each user's activity on the network to better detect and track Advanced Persistent Threats (APT) - threats that may be going undetected in the systems. Pivotal’s data science team works to build customized user behavior models, which are then run against the security data repository to assess individual users, job functions, servers, and server criticality for threat potential. The results are scored and surfaced to forensic experts for remediation. The feedback provided by forensic experts is flagged and sends security alerts via an application or dashboard that is then fed back into the system of predictive models. The models learn from this feedback, continuously improving in accuracy to better identify true security threats and reduce false alarms using scalable machine learning algorithms.