Google Workspace Agents

Google is turning Workspace into a system that can understand your work across Gmail, Docs, Sheets, Chat, Calendar, and Drive, then act on it with Gemini. (workspaceupdates.googleblog.com) Google announced Workspace Intelligence on April 22 at Cloud Next 2026, and said it gives Gemini a real-time view of a user’s work so people no longer have to paste the same context into every prompt. Admins can control which data sources it can use in the Admin console. (workspaceupdates.googleblog.com) The same day, Google introduced Gemini Enterprise Agent Platform, a new version of Vertex AI for building, scaling, governing, and optimizing agents, with tools for integration, DevOps, orchestration, and security. Google said future Vertex AI services and roadmap updates will be delivered through that platform instead of as a standalone service. (cloud.google.com) In plain terms, Workspace Intelligence is the context layer and Agent Platform is the control layer. One gathers the emails, files, schedules, and project history an employee is allowed to see; the other gives information technology teams a place to build and manage fleets of agents that use that context. (workspaceupdates.googleblog.com) (cloud.google.com) Google is also wiring that context into specific products. In Chat, “Ask Gemini” can generate documents and slides, search files from descriptions, find meeting times, assemble daily briefings, and connect to tools including Asana, Jira, and Salesforce; in Docs and Slides, Gemini can use business data and company templates to draft materials; in Sheets, Google says prompt-based filling can be 9 times faster than manual entry. (9to5google.com) (techcrunch.com) The selling point is not just a better chatbot. Google said Workspace Intelligence respects existing user permissions, grounds answers only in content a user can already view, and does not use customer data to train generative artificial intelligence models or for advertising. (workspaceupdates.googleblog.com) That puts the competition inside companies on a different battleground. Google is pitching control over data sources, identity, memory, orchestration, and security as the core product, while also offering access to more than 200 models through Model Garden, including Google models and third-party models such as Anthropic’s Claude. (cloud.google.com) Google paired that pitch with scale numbers at Cloud Next: nearly 75% of Google Cloud customers now use its artificial intelligence products, and 330 customers processed more than 1 trillion tokens each over the past 12 months. Those figures help explain why Google is packaging workplace assistants and agent infrastructure as one enterprise stack. (blog.google) Workspace Intelligence is rolling out now, with Google saying it is on by default for supported editions and began full rollout on April 22, though exact timing depends on each customer’s release track. The immediate question for buyers is less whether Gemini can draft an email than whether Google can become the system that decides which agent gets access to which piece of work. (workspaceupdates.googleblog.com) (workspace.google.com)