Google has rolled out a fresh round of upgrades to Gemini Spark, its always-on AI agent, giving it deeper control over Google Workspace files. The update landed this week, roughly two weeks after Spark’s previous feature drop. It expands what the assistant can do inside Docs, Sheets, and Slides while also making the agent noticeably faster. The change matters for the millions of Google AI Ultra subscribers who already rely on Spark to handle multi-step tasks in the background.
What The New Gemini Spark Update Changes
The headline addition is direct access to editing. Spark can now open and edit Google Docs on a user’s behalf, a capability that previously required more manual back-and-forth. The agent has also gained the ability to edit private spreadsheets and presentations, not just view them, marking a notable jump from its earlier read-only limits.
Expanded Workspace Permissions
Beyond basic editing, Spark now supports a wider range of collaborative actions. These additions focus on shared files and richer content edits rather than simple text changes.
- Reading comments left on spreadsheets and presentations
- Editing shared documents, spreadsheets, and presentations with other collaborators
- Adding images directly into documents and presentations
- Refining and rewriting content inside the Canvas panel, including shared files
These permissions turn Spark from a research assistant into something closer to a full editing partner. A small business owner could ask Spark to update a shared budget sheet after reading a colleague’s comments, then insert a chart image without opening the file manually.
Read more: Microsoft Patches
Gemini Spark Gets A Major Speed Boost
Alongside the new editing tools, Google says the core engine behind Gemini Spark is now more than 50% faster. Long-running research or organizational tasks that once took several minutes should be completed noticeably more quickly. Google credits this improvement to smarter sourcing, a technique that lets the agent retrieve and review multiple reference sources in parallel instead of scanning them one at a time. Google has not published exact benchmark figures, but it frames the change as one of the biggest performance jumps since Gemini Spark’s initial rollout.
This kind of parallel processing matters most for complex requests, such as pulling data from several documents to compile a report. Users working on long-horizon tasks, like synthesizing meeting notes across weeks of emails, should feel the difference most clearly.
Google also cleaned up how Spark communicates while working in the background. The agent now sends more detailed alerts whenever it needs a person’s direct input to keep a task moving forward. That extra context should reduce confusion about why a task has paused.
A new presence filter adds another layer of convenience. If someone is already active in a task’s thread inside the Gemini web app, Spark will automatically suppress duplicate mobile notifications for that same task. This small fix addresses a common complaint about redundant alerts arriving on multiple devices at once.
Read more: Google Pixel 11 Pro
Who Can Use Spark Right Now
This upgraded version of Gemini Spark is rolling out to Google AI Ultra subscribers in most regions where Gemini apps are supported. Google has confirmed that the rollout excludes the European Economic Area, the United Kingdom, Switzerland, and Nigeria, likely due to ongoing regulatory reviews in those markets. Availability elsewhere is expanding gradually rather than all at once, so some Ultra subscribers may see the new tools arrive a few days later than others.
Google AI Ultra remains the only confirmed tier with access to Spark’s full feature set. Josh Woodward, who leads the Gemini app and AI Studio at Google, has publicly teased that Google AI Pro subscribers should expect an access update soon, though no firm date has been announced.
How Spark Fits Into Google’s Agent Strategy
Gemini Spark first launched at Google I/O 2026 as a cloud-based agent capable of working even when a laptop is closed or a phone is locked. It runs on Gemini 3.5 and uses Google’s Antigravity agent harness, allowing it to take autonomous action across Workspace apps under a user’s direction. Since that debut, Google has shipped updates roughly every two weeks, a cadence that suggests the company is treating Spark as a fast-iterating product rather than a one-time launch.
The broader push places Spark alongside a wave of agentic assistants from rival AI labs, including similar background-task tools from Anthropic and OpenAI. Google’s advantage lies in Spark’s tight integration with Gmail, Docs, Sheets, and Slides, apps that hundreds of millions of people already use daily.
Practical Next Steps For Spark Users
Anyone with an active Google AI Ultra subscription should already see the new Gemini Spark capabilities inside the Gemini app without needing to change any settings. For those on Google AI Pro, keeping an eye on official Google channels is the fastest way to learn when broader access opens. Google has not detailed pricing changes tied to this update, and existing subscription terms remain unchanged for now.
