Google Cloud has launched Gemini Enterprise, a new platform it calls “the new front door for AI in the workplace.” Announced at a virtual press event, the offering centralizes Google’s Gemini models, first- and third-party agents, and the core technology that used to be called Google Agentspace into a single agent-driven system designed to make AI agents easier to build and operate across companies.
The product is intended to broaden access to AI-powered agents that automate multi-step workflows and raise productivity across organizations. Thomas Kurian, CEO of Google Cloud, framed the shift in customer activity by saying they were “advancing to build agents” rather than only creating AI-enabled applications.
Gemini Enterprise packages Google’s full AI stack into a unified user experience that brings a no-code workbench to both developers and business users. The platform is organized around six principal elements. The “brains” are Google’s Gemini models, including the newly available Gemini 2.5 Flash Image. The “workbench” refers to the agent creation and orchestration tooling that evolved from Agentspace, giving users a point of control for building, configuring, and running agents. A “taskforce” supplies prebuilt Google agents for specific duties, among them the new Code Assist Agent and the Deep Research Agent.
Deep ties to corporate data are part of the design, with new connectors for Microsoft Teams, Salesforce, Box, Confluence, and Jira so agents can access the right inputs at the right time. Kurian described the personalization layer this way: “We remember who you are and what you do and use it to personalise the context you have when we work with a large language model.” A central “governance” framework sits over all of it so organizations can monitor, secure, and audit agents from a single dashboard, with protections such as Model Armor included. Google says the platform also rests on an open ecosystem of more than 100,000 partners.
During a live walkthrough, Maryam Gholami showed a hands-on example meant to illustrate how the tools operate in real workflows. Gholami noted that the interface mirrors the Gemini experience but is optimized for enterprise processes, with controls that let administrators switch data sources on or off as required. She demoed a custom “campaigns agent” that combined four distinct agents to run market research, produce creative assets, coordinate team communications, and handle inventory actions. The system detected a trend toward sci-fi themes, flagged a 25 percent inventory shortfall, created a purchase order in ServiceNow, drafted an email to store managers, and produced social media assets for promotion.
“Gemini Enterprise is more than just a chat interface,” Gholami said. “It’s an end-to-end AI system that unifies your data, your tools, and your teams, turning weeks of complex work into a single, streamlined conversation.”
Executives from early adopters described business results and cultural shifts tied to agent deployment. Nirmal Saverimuttu, CEO of Virgin Voyages, stressed that “any major disruption like AI requires a cultural transformation to be successful.” He made clear the technology is intended to assist staff rather than supplant them, saying, “Our people are our biggest asset. AI. And never replace our people.” He added, “To me, AI is about getting the best from our people. It’s about unleashing human potential.”
Virgin Voyages has rolled out a fleet of more than 50 specialized agents around the company. The first agent, nicknamed “Email Ellie,” accelerated content production by 40 percent and helped deliver a 28 percent year-over-year lift in July sales. Operational changes included a reported 35 percent cut in agency dependency costs and greater creative independence for in-house teams.
Macquarie Bank is another early user. The Australian bank deployed Gemini Enterprise to every employee and reports that 99 percent of staff have finished generative AI training as part of the rollout.
Google has positioned Gemini Enterprise as an open platform and named partners such as Box, Salesforce, and ServiceNow, each of which announced compatibility for agents built on the system. A new agent discovery tool is planned to help customers find validated solutions from the partner network.
To back adoption, Google launched Google Skills, a free learning hub that offers 3,000 courses, and introduced the Gemini Enterprise Agent Ready (GEAR) program, an educational sprint aimed at helping one million developers build and deploy agents. The company said the platform will be available worldwide in every market where Google Cloud sells products.
Pricing for the new tiers was disclosed at the event. Gemini Business, intended for small firms, starts at $21 per seat per month. Gemini Enterprise Standard and Plus, aimed at larger organizations, begin at $30 per seat per month.
Kurian returned to the theme of widening access as the session closed, stating, “Gemini Enterprise technology is really about reimagining a super powerful AI technology [for the workplace] but making it super easy to use and putting it in the hands of every company and every user in those companies,” Kurian concludes.