Google Cloud: agentic security

Google Cloud used its Next conference on April 22 to push AI agents deeper into security work, adding new security agents and new controls to govern what those agents do. (cloud.google.com) An AI agent is software that can take actions across systems, not just answer questions, and Google’s new Gemini Enterprise Agent Platform is pitched as the place to build, scale, govern, and optimize those agents. Google said the platform adds agent integration, DevOps, orchestration, and security features on top of Vertex AI. (cloud.google.com) Inside the Gemini Enterprise app, Google said customers get an Agent Designer and an Inbox for managing agent activity, while the app itself includes “complete agent observability and traceability” and “granular control.” Those are the mechanics enterprises use to see what an agent did, who approved it, and where it connected. (cloud.google.com 1) (cloud.google.com 2) On the security side, Google said it introduced three customer-facing agents in preview after testing them internally: a Threat Hunting agent, a Detection Engineering agent, and a Third-Party Context agent that is slated for later release. The company said those tools are meant to search for novel attack patterns, find coverage gaps, and enrich workflows with outside data. (theregister.com) Google tied those agents to what it calls “Agentic Defense,” a package that combines Google Threat Intelligence and Google Security Operations with Wiz’s cloud and AI security platform. Google’s Next materials said the aim is to cover code, cloud, runtime, multicloud, hybrid, and AI environments in one stack. (cloud.google.com 1) (cloud.google.com 2) The company paired the product launch with money for the channel. Google Cloud said it is creating a $750 million fund for its 120,000-member partner ecosystem to support agentic AI development, adoption, and education. (prnewswire.com) Google’s partner blog said the fund can pay for software companies to build agents on the Gemini Enterprise Agent Platform, for forward-deployed engineers to work with large systems integrators, and for training and go-to-market support. Google also said partner-built agents will appear in the Agent Gallery inside Gemini Enterprise. (cloud.google.com 1) (cloud.google.com 2) Boston Consulting Group was one of the firms attached to the rollout. On April 22, BCG and Google Cloud said they were expanding their partnership to move customers beyond AI pilots and into enterprise-wide agent adoption with “measurable bottom-line impact.” (bcg.com) Google is making the security case while adoption numbers climb. The company said nearly 75% of Google Cloud customers now use its AI products, 330 customers processed more than 1 trillion tokens in the last 12 months, and direct API use rose to more than 16 billion tokens per minute from 10 billion last quarter. (cloud.google.com) (blog.google) That scale is why governance keeps showing up next to automation. Google has separately warned about “shadow agents,” argued that agent deployments need the same rigor as human-managed accounts and cloud infrastructure, and published audit controls for AI workloads; its documentation also points customers to audit logs in both Google SecOps and Google Cloud. (cloud.google.com) (cloud.google.com) (docs.cloud.google.com) The pitch at Next was straightforward: let agents do more of the routine defense work, but keep humans, logs, and policy controls in the loop. Google Cloud chief operating officer Francis deSouza told reporters the company’s model is “an agentic fleet” that works at machine speed and is overseen by humans. (theregister.com)