Patented AI-Powered Threat Detection Platform Enhances Real-Time Email Security
Detecting, Characterizing, and Remediating Email-Based Threats in Real Time
U.S. Patent No.11,824,870
Abnormal has developed a patented, AI-native platform that proactively detects and stops email-based threats as they happen. This breakthrough technology uses advanced machine learning to understand how people across an organization typically communicate, then flags anything that seems out of the ordinary. By focusing on human behavior instead of static rules, Abnormal can uncover threats that traditional security tools often miss.
This diagram shows how email threats are detected using a system made up of the customer’s enterprise network and Abnormal’s threat detection platform. The platform monitors communications across the network to identify potential risks."
The Patented Abnormal AI Advantage
What makes the AI in this patented system unique is its ability to continuously learn from past communications and apply that knowledge to new messages as they arrive. By analyzing behavioral traits—like how often a sender identity or email address is used—the AI can detect subtle anomalies that might indicate a threat.
Why This Matters
Traditional email security measures often rely on static rules and known threat signatures, which can be insufficient against evolving attack vectors. Abnormal Security's patented approach introduces a dynamic, behavior-based detection mechanism that adapts to new threats, providing a robust defense against sophisticated email attacks.
Meet One of the Inventors
Evan Reiser
Evan Reiser is the CEO and Co-Founder of Abnormal AI, where he leads the company’s strategic vision and innovation in behavioral AI. With over a decade of experience building enterprise-grade machine learning systems, Evan has consistently pushed the boundaries of what's possible in cybersecurity.
Recognizing that traditional security tools fail to detect advanced social engineering attacks, Evan pioneered a patented system that uses behavioral modeling to assess risk in real time. This system—now protected under U.S. Patent No. 11,683,284—identifies security threats by measuring deviations in communication behavior, such as unusual sender identity or email address usage. It marks a significant advancement in using ML to detect identity-based threats like email account compromise and impersonation.
Under his leadership, Abnormal’s AI-native platform has set a new standard for threat detection and prevention—contributing to the company’s recognition as a Leader in the 2024 Gartner® Magic Quadrant™.
How Abnormal's Patented AI Detection Engine Works
- 1. Collecting Behavioral Data: The platform leverages a secure API connection to collect past messages received by employees across an organization.
- 2. Building a Machine Learning Model: A machine learning model is trained on past messages to learn how an employee or company typically communicates.
- 3. Assessing Deviations: When a new message comes in through an API, the system uses its trained machine learning models to analyze or and score the message based on how much it deviates from the employee’s usual behavior.
