Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

The most pleasant agentic AI system is the least secure one. That's the trap Rob van der Veer, Chief AI Officer at Software Improvement Group, lays out clearly: the agent that never says "permission denied" is the agent everyone loves to use. It always works. So we open up all the privileges just to keep it that way. He says we should do the contrary.

Today we're speaking with Rob van der Veer, Chief AI Officer at Software Improvement Group, about how organizations can build trustworthy AI in an era of rapidly evolving technology and regulation — AI security, threat modeling, international standards, and the new challenges posed by agentic AI.

Frontier AI has armed adversaries with unprecedented speed, finding and exploiting vulnerabilities faster than humans can respond. To keep pace, defenders need agentic AI operating on defense. In this demo, see how a Frontier AI Exploitation Defense Agent in Charlotte AI AgentWorks identifies and prioritizes critical risk, then hands off to Falcon for IT Risk-Based Patch Management for controlled remediation, all within the CrowdStrike Falcon platform.

On this episode of Masters of Data, we take you inside the Dojo AI demo we're bringing to the show floor at Black Hat. We walk through the SOC Analyst Agent, Mobot, and MCP back to back. SOC Analyst Agent triages every tier one alert down to the one that actually matters, and Mobot picks up from there, running the investigation in plain English to track down other phishing victims and lateral movement. We also show how MCP pulls that same insight into Claude or Slack. SOC leads tired of alert fatigue and analyst burnout will want the numbers here: 100% of tier one alerts triaged automatically, and 25 hours a week back per person.