Google supercharges Workspace with inbox cleanups and slide-to-video tools

Staring at a hopelessly bloated inbox or digging through five-year-old Drive folders just to answer a routine client email is a universal workplace headache.

At Google I/O 2025, Google rolled out a massive wave of Gemini updates across Workspace. Alongside the reveal, the company shared that Workspace is already dishing out 2 billion AI assists every single month. The new tools are squarely aimed at killing everyday office busywork.

Cleaning your inbox in plain English

Gmail is getting three updates that might actually give you hours back. The first is personalized smart replies. Instead of generic suggestions, Gemini now searches your past emails and Google Drive files to grab exact details, drafting replies that match your usual tone whether you write formally or casually.

Inbox management is also ditching clunky manual filters. You can now tell Gemini in plain text to delete or archive specific clusters of mail, like wiping out all unread newsletters from a specific brand last year. When Gemini spots that you are trying to coordinate with external clients, it will automatically surface a quick booking tool right inside the thread.

Over in Google Docs, a new grounding feature is rolling out today. You can link specific slide decks, reports, or data files directly into a document, and Gemini will restrict its writing assistance purely to those trusted materials. That should stop the model from hallucinating outside facts into your research summaries or business plans.

Slides turn into videos, avatars take the mic

Google Vids is seeing a major expansion. In coming months, users will be able to take an existing Google Slides deck and convert it into a complete video with automated scripts, voiceovers, and animations.

If you do not have the budget for a video shoot or simply do not want to be on camera, Vids is introducing AI avatars in Labs next month to present your scripts. The app is also adding a transcript trim tool to cut out filler words like um and ah, alongside sound balancing and Workspace-wide image generation powered by Imagen 4.

Google Meet also showcased something wild: low-latency speech translation that preserves the speaker's original voice, tone, and inflection. It is kicking off in beta today for Google AI Pro and Ultra subscribers with English and Spanish, aiming to make multilingual calls sound natural rather than robotic.

Why it matters

We think this batch of features hits the sweet spot between utility and zero learning curve. Tools like Docs source-grounding and fast calendar scheduling do not require fancy prompt engineering. They simply cut out ten browser tabs worth of copy-pasting, making them immediately worth using the moment they land in your tenant.

For the broader productivity race, Google is leveraging its unified cloud ecosystem to pressure both Microsoft Copilot and standalone video startups. Turning static slide decks into full videos with synthetic presenters essentially commoditizes the basic explainer video workflow that niche SaaS apps have been charging premium subscriptions for.

There is, however, a clear risk in handing destructive powers to an AI agent. Asking Gemini to mass-delete emails based on natural language commands demands absolute precision. If the model misinterprets an ambiguous query and scraps an old legal thread, the saved minutes will not mean much. We suggest letting Gemini tackle promotional spam first while keeping critical archives strictly out of reach.