Google expands Gemini Live with voice-driven “agents” for daily tasks

Google is upgrading Gemini Live to do more than answer questions through conversation. In a new release, the company says Gemini Live can now carry out productivity workflows through voice, including organizing day plans, handling email, and kicking off multi-step tasks that run in the background.

A central part of the update is Personal Intelligence, which the company describes as Gemini’s ability to connect the dots across past conversations and Google apps such as Gmail, Photos, Search, and YouTube. Google says the assistant can use those connections to provide more helpful answers, including retrieving details from previous chats and connected services.

Google is also introducing a feature called Daily Brief inside Gemini Live. The company says the Daily Brief is a spoken summary that pulls together key updates from Gmail and Calendar, aimed at answering the practical question of what to focus on that day. Users can invoke it through voice, with Google pointing to the prompt “What’s my daily brief?”

On the hands-free email front, Google says Gemini Live now acts as a conversational partner for sorting messages from Gmail. Users can ask natural-language questions such as “Any new emails?” or “Are there any urgent emails from the kids’ school today?” and Gemini can respond by searching, summarizing, starring, archiving, or deleting messages using voice controls.

Google also described agentic task execution through an integration with Spark 1. Spark, as Google explains it, works across Google Docs, Sheets, Drive, and the web, and it can remember goals as it goes. The company says Spark can handle long-running tasks scheduled over days or weeks, even when the user is not actively using the app.

Gemini Live is designed to route a user’s request to the right capability inside a single, unbroken conversation, according to Google. The company provided an example where a user asks what their morning schedule looks like, then follow-ups by requesting event invites from email, creating family calendar events, and adding driving times. Google says Gemini can hand the broader multi-step work to Spark to run in the background while the user continues with their morning.

Google said the upgrade is focused on faster, voice-first workflows, pointing to a usage stat that 63% of users talk to Gemini out loud. The company frames the improvements as a way to translate spoken prompts directly into outcomes rather than keeping the interaction limited to general chat.

As for access, Google said users need to connect apps in their Gemini Personal Intelligence settings and then tap the Live icon in the Gemini app to begin speaking. The company also tied subscription levels to specific features: Spark is available with a Google AI Pro subscription or higher, while Daily Brief is available with a Google AI Plus subscription or higher.

Still, not all operational details are spelled out in the announcement. Google does not specify rollout timing by device, language availability, or how Gemini confirms permission and boundaries when actions involve email deletion, archiving, and calendar updates. It also does not provide technical documentation on how the voice agent decides when to invoke Spark versus other built-in tools.

For Alphabet and the broader AI assistant market, the updates suggest Google is pressing toward practical “do-work” assistants that can persist across time and coordinate multiple Google services. What to watch next is whether Gemini Live’s agentic behaviors expand further across more app categories, and whether Google publishes clearer controls and transparency features for users when assistants take actions on their behalf.