How AI-Based Red Teaming Agents Can Reduce Cyber Risk

Armadin's Evan Peña of on Continuous Testing of Code, Configuration and Identities Michael Novinson (MichaelNovinson) • March 30, 2026

Organizations need to know their controls work before attackers exploit weaknesses. That means security teams should continually test network architecture, configurations and detection systems under real conditions, said Evan Peña, co-founder and chief offensive security officer at Armadin.

See Also: AI Impersonation Is the New Arms Race—Is Your Workforce Ready?

Traditional pen testing and red teaming fall short because of scale, time and expertise challenges, he said. AI attack agents can help by enabling continuous, full-coverage assessments across environments. Agents operate without interruption and evaluate new code, configuration changes and identity risks as they emerge, he said.

"You don't want to wait for a breach to determine whether your security controls or your network architecture configurations are secure or hardened. You'd rather battle-test that to ensure whether it ...