Agentic AI
,
Artificial Intelligence & Machine Learning
,
Black Hat

Ent CEO Lou Manousos on User Intent, Local AI and Machine-Speed Defense

Michael Novinson (MichaelNovinson) •
August 13, 2026    

Lou Manousos, co-founder and CEO, Ent

Artificial intelligence agents are redefining the enterprise endpoint as autonomous processes gain local access to applications, data and tools, requiring security teams to monitor the workflow of both human activity and agent behavior, said Lou Manousos, co-founder and CEO of Ent.

See Also: AI Phishing Now Frighteningly Normal, Hard to Detect

Tasks that take minutes for humans to complete can take only seconds for agents. That speed creates risk when an agent misreads a user’s intent, accesses sensitive data or acts beyond its assigned task, Manousos said. Legacy endpoint tools often see an agent as another process and miss the decisions occurring inside its container.

“Without that type of sub-second response time, the damage may already be done. So, trust is the critical element, and then also having an architecture that can support real-time preventative defense,” he said.

In this video interview with ISMG at Black Hat USA 2026, Manousos also discussed:

Why defenders need telemetry that models user and business intent;
How agentic AI moves enterprise activity beyond SaaS applications and traditional browsers;
How local models and dedicated hardware can support real-time preventive defense.

Manousos has founded and led multiple cybersecurity and enterprise technology companies. He previously was CEO of RiskIQ, before its acquisition by Microsoft, where he led Microsoft Copilot for Security. His expertise spans AI, cybersecurity, enterprise software and national security.