Dive Brief:
- Google is reimagining the human-agent relationship with its newly launched Gemini agent, unveiled Thursday during its Gemini at Work 2026 conference. A unified agentic AI service, Gemini operates in the background, across systems and devices, rather than as a point solution for summaries and enterprise search.
- The agentic iteration of the company’s Gemini tool runs in the cloud and can retain key context for the tasks it performs. It has the ability to create sub-agents to tackle more complex tasks. The Gemini agent is available today in preview to a select set of customers, the company said in an email, with wider availability coming soon.
- “Today, Gemini becomes an agent,” Thomas Kurian, Google Cloud CEO, said during the conference in Mountain View, Calif. “You give it objectives, not just instructions. It plans the work, uses custom skills and tools, connects to your systems, and brings back something finished.”
Dive Insight:
The rapid development of agentic capabilities has left CIOs and other senior leaders grappling with tool sprawl, as platforms add an increasing amount of features.
Google has been working to present a unified hub of its AI capabilities, even rebranding previous versions of its tools under the Gemini moniker. Now, with the agentic tool set in one place, the provider is aiming to simplify adoption for enterprise users.
“Google talked about Gemini Enterprise as a ‘front door’ for AI a year ago, but what sat behind that door was a set of tools employees still had to assemble themselves: models, a no-code agent builder and a catalog of agents,” said Nick Patience, VP and practice lead of AI Platforms at The Futurum Group, in an email to CIO Dive. “This release puts a single agent behind the door that does the assembly.”
The updates announced Thursday also address governance concerns, Patience said, effectively giving agents their own identity within IT ecosystems and their own sandbox, in addition to per-project spend gaps to help bring costs under control.
But the Gemini platform updates fall short of addressing Google’s issues with fragmented branding, said Joe Mariano, senior director analyst at Gartner.
“They’re saying that the Gemini agent is omnipresent across both Google Workspace and Gemini Enterprise, so this raises new questions about where Workspace ends and where Gemini Enterprise begins,” said Mariano. “Why would current clients of both Workspace and Gemini Enterprise need one of the other tools at this point? I don’t know if Google’s been clear enough on that yet.”
Still, the updates to Gemini do indicate a shift in how works get done, according to Mariano, as Google moves away from productizing agents as assistants toward a more focused approach on task completion. The additions of MCP server access and agent building capabilities to Gemini Workspace are also notable, he added.
For CIOs tracking the evolution of the provider landscape, a close focus on the fundamentals of AI still matters, according to Patience.
“CIOs should go in with their eyes open,” Patience said. “Model choice is open, but the memory, context and skills an agent builds up over time live in Google’s layer, and that’s where the real switching costs will sit.”
