Google revamps Gemini Enterprise for the agentic era: what’s new

At Google Cloud Next, the company’s annual enterprise conference held on Wednesday, the tech giant unveiled a sweeping overhaul of its enterprise AI suite, signaling a clear pivot toward the agentic era. As organizations increasingly deploy autonomous AI agents to handle complex workflows, the challenge of managing them securely and at scale has become acute. Google’s answer is the new Gemini Enterprise Agent Platform, a unified environment designed to simplify the entire lifecycle of agent development, deployment, and governance.

Evolving directly from Vertex AI, the Agent Platform consolidates the model selection, building, and tuning capabilities that customers already rely on, while adding fresh layers for agent integration, security, DevOps, and orchestration. Google CEO Thomas Kurian emphasized that the platform “brings together the model selection, model building, and tuning services of Vertex AI that customers love, along with new features for agent integration, security, DevOps, orchestration, and more.” The revamped Gemini Enterprise now offers access to over 200 models, including Google’s own Gemini 3.1 Pro, Nano Banana 2, and Gemma open models, as well as competitive offerings from Anthropic, such as its newly released Opus 4.7. Notably, all Vertex AI services will now flow exclusively through the Agent Platform.

For developers, the platform provides a full-stack environment for building agents from scratch. They can design, scale, and govern an agent’s entire lifecycle within a single interface. Features like MCP support and an upgraded Agent Development Kit allow developers to structure agents into sub-networks, enhancing their reasoning capabilities and enabling them to tackle multi-step tasks more effectively. Google also highlighted improvements such as a faster runtime and a Memory Bank, which help agents delegate tasks to one another and maintain context over longer sessions. “Gemini Enterprise is now an end-to-end system for the agentic era, built for agents that can execute complex, multi-step work processes,” the company stated.

Security remains a top concern as autonomous agents proliferate. Google has baked in Agent Identity, a system that assigns each agent a unique cryptographic ID. For organizations that want to test before deploying, the Agent Simulation tool allows teams to “stress-test your agents against real-world scenarios before they ship.” Once agents are ready, developers can publish them directly to the Gemini Enterprise app, where employees can either run those pre-built agents or create their own using no-code or low-code tools like Agent Studio and Agent Designer.

A live demonstration at the conference showed how users could deploy multiple agents simultaneously within the enterprise app to tackle an inventory or marketing challenge. Each agent handled a specific part of a multi-step project for a furniture company, pulling relevant data and strategy points from the organization’s Workspace contents. This approach, Google argues, turns the app into a collaborative team of digital workers.

Running multiple autonomous agents introduces significant privacy and security risks, especially when non-developer employees are involved. Google emphasized that the new Gemini Enterprise simplifies guardrails and permissions before users can access any agent. The company said it “provides the same level of oversight and auditability found in essential business applications like payroll or quarterly financial reporting.” The Gemini Enterprise app sits atop the Agent Platform, which Google described as a single control plane for governance. “Both no-code and pro-code agents are managed through a consistent model for identity, security, and auditing,” the company added.

Beyond the Agent Platform, Google announced Agentic Data Cloud, a new data architecture designed to help scale AI agents. This feature allows developers to query data instantly without moving it out of AWS or Azure, leverage data science tools across multiple surfaces, and enrich files with metadata to give agents more semantic context.

At the Workspace level, Google introduced Workspace Intelligence, which uses Gemini reasoning to understand “complex semantic relationships within your Workspace apps (such as Docs, Slides, or Gmail) content, your active projects, your collaborators, and your organization’s domain knowledge.” While this may sound similar to existing Gemini capabilities, Google framed Workspace Intelligence as an additional tool that Gemini will leverage when automating tasks like slide generation and project prep. The new feature includes proprietary infographics in Docs and advanced personalization tailored to a user’s style. “Workspace Intelligence retrieves your relevant emails, chats, files, and information from the web to transform ideas into professionally formatted drafts that mimic your exact voice, brand, style, and company templates,” Google said.