Google Cloud Security has outlined its vision of an agentic SOC (Security Operations Center) and announced wide-ranging enhancements across its unified cloud security platform, aiming to make security an enabler of AI-driven business growth.
At the Google Cloud Security Summit 2025, the company showcased its Alert Investigation Agent, now in preview. Built on frontline Mandiant expertise, the AI-powered agent can autonomously enrich alerts, analyse command-line activity, and generate recommended next steps – drastically reducing manual workload for security teams.
Google also expanded its Unified Security and Trusted Cloud offerings:
-
Compliance Manager (preview) – Streamlines AI-specific policy enforcement and evidence generation for audits.
-
Agentic IAM (coming soon) – Auto-provisions agent identities across runtimes, enabling secure agent-to-agent interactions.
-
Expanded Data Protection – Sensitive Data Protection now covers AI workloads in Vertex AI, BigQuery, and CloudSQL, alongside image scanning for sensitive data.
-
Network Security Upgrades – Including Cloud NGFW for AI HPC workloads and enhanced Cloud Armor enterprise policies.
Mandiant Consulting further announced AI security governance and adversarial testing services, addressing rising demand for AI-specific risk frameworks.
“These innovations demonstrate our commitment to helping enterprises adopt AI responsibly while defending against evolving threats,” said Ramsey. “By using AI to empower defenders and securing AI from the ground up, Google Cloud is building the foundation for the future of cybersecurity.”
