Google launches Gemini Spark, a 24/7 autonomous personal AI agent, in India

Google has introduced Gemini Spark, an autonomous personal AI agent built on Gemini 3.6 Flash, to users in India, expanding a feature it first previewed at Google I/O 2026 into general availability in one of its largest markets.

What's new

According to Google's announcement, Gemini Spark "works quietly in the background - even when your laptop is closed or your phone is locked," positioning it as an agent that keeps handling tasks rather than waiting for a user to open an app. It "connects natively with your Google Workspace tools like Gmail, Docs, and Sheets with zero setup required," and Google describes it as functioning as "an active partner that gets real work done for you, even while you're sleeping."

Google says the agent is rolling out first to Google AI Ultra subscribers in India, with Google AI Pro subscribers gaining access "over the next few weeks." On guardrails, Google says the agent is "designed to ask you first before performing high-stakes actions like spending money or sending emails," and that users control both whether to enable it and which apps it can connect to.

Reported use cases include organizing email in a user's own writing style, summarizing incoming RSVPs, planning multi-service travel itineraries, and tracking recurring subscriptions — the kind of multi-step, multi-app busywork Google is positioning agentic AI to absorb rather than simply assist with.

Context

Spark was first shown as a preview at Google I/O 2026; this release marks its move into a live regional rollout rather than a demo. It arrives on the heels of Google's broader push this year to bring Gemini 3.5 and 3.6 Flash to general availability across its consumer and developer surfaces, and it lands in India specifically — a market Google has repeatedly prioritized for Gemini distribution given its scale and mobile-first user base.

Why it matters

Spark is a clear step past chat-and-respond assistants toward agents that act on a standing basis, inside a user's existing accounts, without being re-invoked for every task. The explicit approval gate on "high-stakes" actions — spending money or sending mail — signals Google is trying to thread the needle between genuine autonomy and the trust problems that come with letting an AI agent operate unsupervised inside email and documents. How that gate holds up in practice, and whether Spark expands beyond India on the same terms, will be an early signal for how far consumer-facing agentic AI can go before users start pulling back the leash.