Enhanced AI Classification Engine in Cofense Triage learns from analyst decisions and enables automated processing of high-confidence email classifications
LEESBURG, Va., November 12, 2026 – Cofense, the leader in AI-driven post-perimeter phishing defense, today announced enhancements to its Phishing Defense Platform with an updated AI Classification Engine (AICE) within its Phishing Remediation solution. AICE uses customer-trained AI to classify reported emails based on an organization's unique categories and enables automated processing of high-confidence decisions. This helps security teams reduce manual triage, accelerate phishing response, and free analysts to focus on complex threats that require human expertise.
Unlike static AI models, Cofense’s enhanced AI combines the power of its globally trained phishing detection models with AI trained on an organization’s own analyst decisions and threat environment. Together, these capabilities enable faster, more accurate threat classification and unlock greater automation of phishing response, reducing manual effort while allowing security teams to focus on the threats that matter most. This advancement reinforces Cofense’s commitment to delivering intelligent, adaptive, and automated phishing defense.
“AI should help analysts spend less time on repetitive email handling and more time on the threats that need their expertise,” said Jason Meurer, Director of Product Management at Cofense. “That requires AI that adapts to how their organization works and unique profile of threats they see. By learning from analyst decisions, explaining its classifications and letting teams set their own automation thresholds, AICE gives security teams the confidence to automate more of the response process while staying in control.”
Available within Cofense Triage, the enhanced AI Classification Engine advances phishing response in three areas:
- Greater automation and analyst efficiency. High-confidence AI classifications trigger automated email processing, reducing manual triage and helping teams focus their expertise on complex, high-risk threats.
- Adaptive AI tailored to each organization. Customer-trained models learn from analyst decisions and can be retrained on demand, delivering more precise classifications and adapting to evolving threats and workflows.
- Explainable AI with security teams in control. Transparent AI decisions and configurable confidence thresholds give teams control over automation, while data and processing remain securely within the customer’s Triage environment.
In testing and validation, high-confidence AI classifications automated at least 30% of email handling work, with accuracy exceeding 98%.
Triage also provides model performance metrics, automation coverage and estimates of analyst hours saved, helping organizations evaluate results and refine their use of automation.
“Our investment in AI is focused on helping security teams automate phishing defense without increasing manual work,” said Marc Olesen, Chief Executive Officer of Cofense. “Customer-trained AI models, explainable decisions and analyst-controlled automation give organizations a way to increase the speed and scale of response while keeping human expertise focused where it matters most.”
Additional Resources
- Learn more about Cofense Phishing Remediation: https://www.cofense.com/pdr-platform
About Cofense
Cofense is the leader in post-perimeter phishing defense, built for the reality that phishing gets through. Cofense helps enterprises identify threats already in employee inboxes, remediate active attacks fast, and reduce future risk. The Cofense platform combines phishing-specific AI with expert validation, automation, and collective insight drawn from millions of real-world user interactions. Human-supervised intelligence improves accuracy, eliminates guesswork, and accelerates response so security teams can act with confidence. Global enterprises including Mastercard, Accenture, and Toyota trust Cofense to strengthen organizational resilience against the most persistent cyber threat: phishing.