Google is making significant strides in the realm of artificial intelligence (AI), particularly with the launch of its Gemini Enterprise Agent Platform. This platform serves as a centralized hub designed to manage workflows for agentic AI, which includes both Google-developed and third-party AI agents. The introduction of this platform, just months after the debut of Gemini Enterprise, signals Google’s commitment to enhancing the capabilities of AI in professional settings.
The Gemini Enterprise Agent Platform is not just a management tool; it is a comprehensive solution that integrates a variety of existing and new functionalities. A standout feature is the ability to assign each AI agent a unique cryptographic ID. This ID serves a crucial role, as it is the reference point for every action an agent undertakes. According to Thomas Kurian, CEO of Google Cloud, these agent IDs are crafted to adhere to “defined authorization policies that are traceable and auditable.” This aligns with the growing demand for accountability and transparency in AI operations.
At the recent Google Cloud Next 26 conference, held in Las Vegas from April 22 to April 24, Kurian emphasized the incorporation of “zero trust verification” into every agent and at every orchestration step. This is particularly pertinent in an era where the need for secure digital operations is paramount.
### Addressing Unique Identity Risks
The advent of AI agents is set to revolutionize identity management, particularly for security professionals. Traditional non-human identities, such as application programming interface (API) keys or service accounts, function in a deterministic manner. In contrast, AI agents exhibit autonomy and goal-oriented behavior. They are capable of parsing high-level objectives, breaking them down into actionable steps, and executing these procedures independently across multiple applications to achieve the overarching goal.
This autonomy introduces a new class of dynamic digital entities that operate on behalf of humans and can make independent operational choices. To manage this complexity effectively, agent identities will be cataloged in the Agent Registry—a central library within the Gemini Enterprise Agent Platform that indexes every internal agent, tool, and skill.
Furthermore, the introduction of the Agent Gateway provides a single dashboard for overseeing a fleet of AI agents. This feature allows users to enforce policies governing all agent-to-agent and agent-to-tool interactions. It supports various protocols related to agentic AI, such as the Model Context Protocol (MCP) and Agent-to-Agent (A2A) communication. As a Google Cloud statement articulated, this platform facilitates secure, unified connectivity while enforcing consistent security policies and protections against vulnerabilities like prompt injection and data leakage.
Model Armor, another integral component of the platform, serves as Google Cloud’s defensive layer against adversarial threats, including prompt injection, sensitive data leaks, and harmful content.
Security teams are increasingly challenged by the presence of agents—both authorized and unauthorized—operating within their environments. As Francis deSouza, COO of Google Cloud, noted, managing access control for authorized agents is critical. The dynamic nature of agents necessitates an adaptive approach to permission management, distinguishing them from static human identities.
### New Security Features and Enhancements
A significant announcement at Cloud Next 26 was the introduction of several advanced security features, including Agent Anomaly Detection. This innovative tool utilizes statistical models and a large language model (LLM) framework to detect unusual behavior in real time, thereby identifying potential threats through patterns that might indicate suspicious reasoning.
Agent Anomaly Detection complements existing features like Agent Threat Detection, which monitors for harmful activities such as reverse shells and connections to known malicious IP addresses. Additionally, the newly established Agent Security dashboard aggregates various security functions into a unified platform, streamlining threat detection and risk analysis within Google Cloud environments.
The new dashboard will assist security teams in mapping relationships between AI agents and models, automating asset discovery, and scanning for vulnerabilities across operating systems and language packages. These enhancements build upon the current security tools available through Gemini Enterprise, including Agent Compliance and Agent Policy features.
### A Broader Commitment to Cybersecurity
The launch of the Gemini Enterprise Agent Platform and new security features is just one part of a broader series of announcements made during Cloud Next 26. Google’s acquisition of Israeli cloud security firm Wiz has led to an expansion of its AI-Application Protection Platform (AI-APP), integrating security measures directly into the development workflow. The updates include enhanced real-time vulnerability scanning, AI-generated code security, and automated remediation capabilities across various development platforms.
Google also rolled out three new agents aimed at aiding cybersecurity professionals. The Threat Hunting agent is designed to help teams identify novel attack vectors and stealthy adversarial tactics, while the Detection Engineering agent targets gaps in coverage to improve threat detection processes. Both agents are currently available in a preview mode, and a Third-Party Context agent is set to enrich security teams’ workflows with contextual data from external sources.
In terms of analytics, Google reported that its Triage and Investigation agent processed over five million alerts in the past year, significantly reducing the time required for manual analysis. Among other innovations, Google introduced a new dark web intelligence feature designed to analyze vast quantities of external events with high accuracy, identifying pertinent threats.
As Google charts a course into the future of AI and cybersecurity, it also announced its intent to invest $750 million in a new agentic AI partner fund aimed at supporting global consulting firms and systems integrators in the development and implementation of AI solutions. This fund aims to foster innovation in AI, making Google a pivotal player in shaping the future landscape of agentic AI and cybersecurity.