KnowBe4 is leaning hard into autonomous AI at RSAC 2026, using the conference to spotlight an expanding suite of AI agents it says is reducing security administration from hours to seconds.
The company’s AIDA platform, short for Artificial Intelligence Defense Agents, now includes eight agents after the recent launch of the AIDA Orchestration Agent. That newest addition is fully autonomous: it independently creates, schedules, and manages personalized security awareness training and phishing simulations at the individual user level, adapting to each person’s risk profile without any human administration involved. KnowBe4 says users on the Orchestration Agent reduce their risk score by almost three times compared with non-AIDA customers.
Two other recent additions to the suite: the Deepfake Training Content Agent, which generates custom deepfake training content featuring leaders from within a customer’s own organization, and the Custom SAPA Agent, which delivers security awareness assessments tailored to an organization’s specific policies, security stack, and environment rather than a generic question set. The Custom SAPA Agent draws on more than five years of usage data from 50,000 organizations and five million assessments.
AIDA as a whole harnesses 15 years of user behavior and threat intelligence data. KnowBe4 markets the platform as the only offering in the industry delivering this depth of autonomous human risk management capability, with more agents in development.
“We’re leading the way in AI-powered defense but we’re still only scratching the surface of what autonomous AI can do for security teams,” said Bryan Palma, CEO at KnowBe4. “Our vision is to move security leaders away from reactive, manual processes and toward a state of autonomous, data-backed defense that protects humans and AI agents.”
KnowBe4 is trusted by more than 70,000 organizations worldwide. Its HRM+ platform includes awareness training, cloud email security, real-time coaching, anti-phishing, and the AIDA agent suite. The company has been building AI capabilities for a decade, with AIDA’s predecessor dating back to 2016.