Sublime Attack Score: Explainable, AI-backed threat analysis

Sublime's Attack Score is a sophisticated AI-backed feature designed to evaluate potential email threats by combining the strengths of both black box AI models and rule-based systems. It leverages machine learning to analyze various signals within emails, such as headers, attachment metadata, and link analysis, to identify attack patterns and unusual activities. The system uses an interpretable XGBoost model to ensure transparency and explainability, providing users with clear verdicts and insights into the rationale behind each decision. This approach not only saves manual investigation time but also maintains data privacy by using privacy-preserving feature engineering that excludes personally identifiable information. The platform continuously updates its model with new data and signals to stay ahead of emerging threats, ensuring that security solutions remain effective and trustworthy while empowering analysts with actionable insights.