Reveelium detects abnormal machine and user behavior thanks to its UEBA (User and Entity Behavior Analytics) technology: immediately upon installation, unsupervised learning algorithms identify the behavior of all assets on the system and pinpoint relevant, contextualized anomalies.
The approach combines machine learning and deep learning methods from academic cybersecurity research: graph theory, LSTM-type recurrent neural networks, supervised and unsupervised classification algorithms, and more.

Reveelium technology enables offline log analysis to conduct digital forensics operations or post-incident audits of information system activities. In this "cold log review" configuration, logs can be ingested into a Reveelium lab environment to verify that conventional controls have not missed a cyber threat or deviant behavior indicative of a cyberattack or insider fraud.

Reveelium identifies the origin and path of behavior considered abnormal within the IS, presenting its characteristics. It is capable of identifying alerts that constitute a threat and displaying the decision logic that led to escalating the severity rating of that threat.
Once the threat is proven and recognized, the technology identifies similar behaviors within the information system and can outline potential attack scenarios to protect against them.

Reveelium automatically generates reports to quantify its activity and includes threat-related indicators: affected IP(s), alert counts per threat, criticality indices, histograms, threat risk evolution over time, decision trees, and more.