We’re excited to introduce significant improvements to our detection engine now, enhancing our ability to identify sophisticated email-based threats using more intelligent and context-aware analysis. This update represents a step forward in advanced, LLM-powered detection.
Key Enhancements:
Sender History Context:
We now incorporate richer historical sender data into the LLM to improve classification accuracy. This context helps the model make smarter decisions by more accurately identifying unusual or suspicious senders.Attachment Content Analysis:
We’ve enhanced our analysis of attachment content by going deeper and including the extracted text in multiple file formats (such as PDF, txt, etc.)
This is especially valuable in cases where the email body is empty and only the attachment contains malicious intent, the model now has a fuller context for making decisions.
These updates significantly improve how we identify threats, offering a more intelligent and nuanced layer of defense.