This brief distills five core findings from the 2026 451 Research Report on how agentic security platforms are transforming the SOC. Each finding is anchored in data and organized around a single architectural conclusion: the platform that enables AI returns and the platform that secures them are increasingly the same platform.
For the second year, SentinelOne commissioned 451 Research to survey 611 North American security decision-makers on the state of AI in the SOC. The full report covers platform consolidation trends, AI maturity and sequencing, the data barriers stalling operationalization, the analyst experience shift, and the attack surface AI deployment creates.
In this session, SentinelOne Principal Evangelist Chris Hosking and Senior Manager of Solutions Engineering Carlos Payés cut through the regulatory complexity to give CISOs a clear, actionable picture of what the EU AI Act demands, where enterprises are most exposed, and how visibility and governance over AI usage is now a security imperative, not just a compliance checkbox. You will leave with: A working understanding of how the EU AI Act classifies risk and what deployer obligations actually mean for your organisation A clear view of the shadow AI problem and why you cannot comply with what you cannot see A practical baseline for AI security governance that maps to both regulatory requirements and real-world enterprise risk This is a security leader's guide to acting now, before enforcement begins.
A Conversation with SentinelOne’s Chief AI Officer Join Gregor Stewart, SentinelOne's Chief AI Officer, and Chris Hosking, Principal Evangelist, for a two-part look at what that access means. Gregor takes the frontier view: how fast these models are moving, and what that changes about the assumptions built into today's security model. Chris takes the defender's view: how SentinelOne puts that same intelligence to work protecting your environment right now. What you'll learn: What programs like Glasswing, Daybreak, Trusted Access for Cyber (TAC), and Anthropic's Cyber Verification Program (CVP) actually mean, and which ones are worth signing up for yourself Where Mythos-class and GPT-5.5-class models improve analyst output, and where operational controls still matter How SentinelOne connects that frontier-model access to security governance A look inside the hands-on work