A new wave of phishing attacks is forcing security teams to rethink their defenses, and IRONSCALES believes AI agents are the answer.
The Atlanta-based email security firm this week unveiled its Winter 2026 Release, introducing three specialized AI agents designed to help organizations counter what it calls “Phishing 3.0,” a new generation of AI-powered impersonation attacks.
A-driven threats accelerate as the threat landscape widens
The release comes at a time when AI-driven threats are accelerating.
According to a recent Osterman Research study cited by the company, 88% of organizations experienced at least one security incident in the past year that damaged trust in digital communications.
More than 80% reported increased attacker focus on exploiting trusted channels, and nearly one in five security leaders said awareness training alone is no longer enough.
“Phishing 3.0 is flawless impersonation at scale,” said IRONSCALES CEO Eyal Benishti. “Attackers use AI to research organizations, craft personalized lures, and bypass pattern-based detection on the first attempt. No malicious payloads, just pure social engineering.”
Three new digital defenders, one goal
To counter these “flawless” threats, IRONSCALES is moving from reactive defense to a preemptive strike. The company’s new architecture features three distinct AI agents:
- The Red Teaming Agent mirrors attacker reconnaissance tactics, scanning open-source intelligence such as social media and public filings to identify potential exposure before attackers exploit it.
- The Phishing SOC Agent conducts in-depth forensic analysis of suspicious emails, delivering a detailed security assessment without tying up internal teams
- The Phishing Simulation Agent creates highly personalized phishing simulations based on real-world data, testing employees with scenarios that resemble genuine adversary tactics rather than generic templates.
“Security teams are stretched thin,” said Audian Paxson, principal technical strategist at IRONSCALES, in a statement. “The Osterman data confirms what CISOs already know: the threat has outpaced the tooling. These agents don’t replace your team – they augment what your team can do.”
Expanding protection beyond the inbox
Beyond inbound threat detection, the Winter 2026 Release adds integrated email encryption for outbound messages.
The system uses adaptive AI to automatically apply policy-based or user-initiated encryption, aiming to strengthen compliance and reduce reliance on manual employee actions.
The company also expanded its deepfake protection for Microsoft Teams, enhancing voice detection capabilities and introducing automatic profile learning.
The update allows the system to learn employees’ voice patterns during regular meetings, helping flag impersonation attempts even when cameras are turned off.
Leadership team additions signal further growth
To steer this new wave of innovation, IRONSCALES has bolstered its leadership, bringing on cybersecurity veteran Steven Malone as chief strategy officer and Amit Bluman as SVP of research and development.
With more than 17,000 organizations using its platform, IRONSCALES is betting that preemptive, AI-driven defenses will become essential as cybercriminals increasingly automate their attacks.