How AI is Enabling More Proactive Security

A pulse survey of 44 CISOs, SOC leaders, and practitioners at RSAC 2025 finds SOCs mired in reactive, manual workflows—70% spending ≥75% of their time on reactive tasks and ~60% manually reviewing over half of low-severity alerts—while 84% remain concerned real threats hide in those alerts; traditional automation is seen as rigid and error-prone, with 64% reporting frequent/occasional false positives or unintended actions; respondents see near-term value for agentic AI in triage (64%), threat hunting and correlation (52%), rule tuning (48%), investigations (46%), containment (39%), and generating summaries/orchestrated responses (34%).