Google on Wednesday introduced several new ways for Gemini AI assistant to create and edit content in Workspace apps such as Docs, Slides and Sheets. 

The changes, said Julie Geller, principal research director at Info-Tech Research Group, represent “incremental improvements more than revolutionary features, but they address real workflow gaps. The actual value is that Google is embedding AI assistance directly into the tools people use every day, like Docs, Sheets, and Slides.”

With the “help me create” feature in Docs, Gemini creates a draft based on user instructions, pulling together information from other Workspace files such as documents, chat messages, and emails. It’s accessed via a side panel or the new prompt bar at the bottom of a document. Once a draft is created, Gemini can then make edits to specific sections of text within the document.

Gemini can also match the writing style across a document, which can be particularly useful when multiple coworkers create a document. A “match doc” feature is used to add information to an existing template.

For many office workers, the biggest hurdle to starting a project is manual prep work and “digging through emails to get a first draft on the page,” Yulie Kwon Kim, Google vice president of product for Google Workspace, said in a briefing. “Now, Gemini handles that for you, drawing on information across your Drive, your Gmail, your Chat, to deliver personalized outputs based on your specific context. Once you have a draft you can then co-edit with Gemini live in the document.”

For Sheets, Gemini can take turn raw data in a spreadsheet with raw data into a dashboard with visualizations using natural language prompts. Gemini sets out a series of steps for user approval and asks clarifying questions if needed as it carries out a task. 

Gemini and Sheets are “moving from a tool you work in, to a collaborative partner,” said Kim. “With a single prompt, Gemini pulls relevant data from across your Gmail, Chat, Drive, to create a formatted spreadsheet in seconds,”

Gemini lets users describe what they want in plain English, without needing to know complex formulas. Additionally, a “fill with Gemini” function in Sheets lets users quickly populate cells with relevant information. 

Google cited a company research document as an example. Gemini can determine from the column headers what information is required — HQ location, revenue, and market cap, for instance —and then search the web to generate the information.

While the tool could potentially save office workers time, hallucinations are always a concern with generative AI (genAI), particularly when working with important business data. Google claims Gemini in Google Sheets has achieved a near-human level of accuracy in benchmark tests, and users are given source links to check the data themselves. 

Google

“Accuracy is a top challenge for genAI, and user trust is the ‘final boss,’” said Amy Machado, senior research manager at IDC. “Regardless of the vendor, even the most sophisticated tools will fail to achieve meaningful adoption without it.”  

With that in mind, Google is prioritizing the “context stack,” she said, with a focus on “optimizing document AI, RAG architectures, and vector embeddings that power these new features.

“The goal is high-precision, grounded retrieval and outcomes that earn user confidence,” said Machado.

Gemini in Drive and Slides

There are also updates to Google’s Drive storage. In response to user search queries in Drive, AI Overviews provides a list of relevant files alongside a short summary of their contents. For more in-depth questions, it’s possible to create custom file repositories based on specific files and folders, or even Workspace apps, with Ask Gemini in Drive. 

The Gemini features in Google Drive are the “sleeper story” of the updates this week, said Mike Leone, practice director, data, analytics & AI, at Omdia. “Turning Drive from passive file storage into something closer to an active knowledge-base could change how people think about their own information,” he said.

“I’ve been watching every major platform try to solve the ‘I know this is somewhere in my files’ problem, and Google’s approach of layering AI Overviews on top of Drive search feels like the most natural implementation I’ve seen.”

In Slides, users can edit individual slides with natural language prompts to Gemini, adapting existing design components and text as directed. It’s possible to generate a new slide that matches the same design and formatting across the rest of the deck.

In the future, Google said it will be possible to generate a full presentation from scratch based on Workspace documents.

All of the features roll out this week in beta, and will be available to AI Ultra and Pro subscribers, and Workspace customers in the Gemini Alpha program.

The features streamline processes and could save time on repetitive tasks, said Geller.  “Instead of switching between applications or starting from scratch, employees get support for drafting, organizing data, and summarizing information built right in.”

But she noted “friction points” that could undermine any time-savings. “Ask Gemini function can generate verbose responses without consolidation, leaving you with repetitive content to clean up manually,” said Geller. “And basic things like pasting screenshots for context don’t work the same way as on other platforms, which limits how useful the AI can actually be.”

Leone called the individual features “solid but not groundbreaking, since AI-assisted writing and slide generation have been around for a while now. What’s actually interesting is how Gemini now pulls from Gmail, Drive, Chat and Calendar to build a first draft or assemble a spreadsheet.

“Right-sized context is king right now in enterprise AI, and Google has a natural advantage because so much of a person’s work context already lives inside the Workspace ecosystem,” he said. “I’d expect strong business demand around Sheets and Drive specifically.”

Competing with Microsoft 365 Copilot

Google’s Gemini announcements follow close on the heels of updates to Microsoft 365 Copilot unveiled Monday, including its Copilot Cowork AI agent.

“When it comes to document applications and content creation, Google’s announcements are on par with what Microsoft offers,” said Machado. “Where Microsoft appears to be differentiating is in driving more multi-model and cross-application, integrated agentic experiences.”

Rather than competing directly with Microsoft 365 Copilot, Google is taking a different approach, said Geller. “Google isn’t playing catch up here because they’re moving on a different track,” she said. 

Microsoft is focused on user-facing workflow automation, said Geller, while Google is “offering both lightweight assistants in Workspace and developer tools for building agents. 

“They’re not directly competing yet,” she said. “If these features work reliably and save time, they deepen your investment in the ecosystem, which is exactly what both Google and Microsoft are counting on. 

“Both are iterating quickly, but speed alone won’t determine the winner. The real differentiator will be the way the customer feedback loops are designed, not just shipping features.”

While Microsoft described Copilot as an “agent” in its Wave 3 launch, Google steered clear of the term in relation to Gemini for Workspace. “They frame Gemini as a ‘collaborative partner,’ which tells you where this sits on the AI maturity curve,” said Leone.

“From my recent conversations with Google’s Workspace and Gemini teams, I know their broader agentic vision is much more ambitious, so stay tuned for that. Google has arguably the strongest foundation AI model and agent development frameworks in the market right now, and today’s features show how well Gemini can connect context across the Workspace ecosystem.

“But Microsoft is setting the pace for agentic in the workplace, and that’s the gap Google needs to close.”