Google has introduced Gemini Spark, a new 24/7 personal AI agent designed to handle complex digital tasks for users. This AI agent is built on Gemini 3.5 and leverages the Antigravity harness for background processing. It integrates deeply with Google Workspace tools, including Gmail, Docs, and Slides, enabling it to automate workflows such as parsing financial statements, tracking school updates, and synthesizing meeting notes across platforms. Users can teach Spark new skills, set up recurring triggers, and manage workflows that were previously time-consuming. The AI agent is cloud-based, meaning it remains active even when a device is locked or powered down.
— Josh Woodward (@joshwoodward) May 19, 2026Introducing Gemini Spark!
Our 24/7 personal AI agent designed to proactively manage tasks and help you navigate your digital life, all under your direction.
Coming to trusted testers this week, and as a Beta for US Google AI Ultra subscribers next week! pic.twitter.com/EUjmR4HYdl
Gemini Spark is first rolling out to trusted testers, with a planned Beta launch for U.S.-based Google AI Ultra subscribers soon after. The system is designed to prioritize user consent, requiring confirmation before any high-stakes actions. Google is also expanding Spark’s capabilities through new app connections, with partners like Canva, OpenTable, and Instacart available at launch and more integrations planned. This marks a major evolution for Google, moving from a question-answering assistant to an active digital partner, as it aims to deliver more personalized, task-oriented AI experiences to a broad user base.