As attacker techniques continue to evolve, we’ve introduced two enhancements that deepen credential-phishing and attachment detection, and accelerate vendor fraud analysis and classification.
Enhanced Credential-Phishing and Attachment Detection
Abnormal has enhanced its attachment analysis engine to improve detection of malicious file behaviors across attachment types. These enhancements apply to 15 entity attributes—including QR codes, JavaScript inside PDFs, and password-protected nested archives—and include 21 new behavioral heuristics.
The update also enhances detection for scenarios involving forwarded content hidden inside nested .eml attachments and malicious web archive content embedded within DOCX files, giving the engine significantly more surface area to catch threats before they reach users.
AI Workflow Enhancements for Vendor Fraud Cases
Abnormal has introduced AI-powered enhancements to its automated vendor fraud analysis and classification. The update combines advances in large language model (LLM) analysis, automated sampling, and a deterministic vendor fraud policy layer to automatically generate vendor fraud cases for relevant detections.
The improvements increase overall processing throughput by approximately 38% and generate a greater number of high-confidence vendor fraud cases.
The latest enhancements provide customers stronger protection against evolving email threats and faster detection of vendor fraud cases. No configuration changes required.