Gemini Is Becoming a 24/7 AI Agent—and Google Says 900 Million People Use It Monthly

The chatbot era trained people to open a window, type a request and wait for an answer. Google’s next version of Gemini is built around a more ambitious idea: an AI assistant that anticipates needs and continues working after the conversation ends.

Google says more than 900 million people across over 230 countries and territories now use Gemini each month. That scale gives its shift toward agentic behaviour enormous significance.

From answers to actions

The updated Gemini experience includes proactive daily briefs and Gemini Spark, an agent designed to help complete tasks around the clock. Google is also introducing Gemini 3.5 Flash, which it presents as a faster model suited to responsive, real-world workflows.

An agent is more than a chatbot with a new name. It can potentially plan a sequence of steps, use tools, monitor progress and return with a result. In practice, that could mean researching a trip, organising a schedule, tracking a project or coordinating information across several services.

The attraction is obvious. Many digital tasks are not difficult because of one hard decision; they are exhausting because they require ten small actions across different apps.

Convenience creates new responsibilities

A proactive assistant needs more access than a simple question-answering system. It may need calendar details, location, messages, files and preferences. The more useful it becomes, the more carefully permissions and mistakes must be handled.

Users need to know what an agent can see, what it is allowed to change and whether an action requires approval. Businesses will need audit trails. Platforms will need safe ways to stop or reverse a workflow.

There is also a difference between helpful anticipation and intrusive automation. A daily brief can save time, but an assistant that guesses incorrectly could create noise or make decisions the user never intended.

The agent race is entering the mainstream

With hundreds of millions of monthly users, even a gradual rollout can change public expectations quickly. People may stop judging AI only by the quality of its written answers and start judging it by reliability: Did it complete the task? Did it ask at the right moment? Can I trust what it changed?

The winning assistant may not be the one that appears most intelligent in a single conversation. It may be the one that becomes dependable enough to handle ordinary life without creating new work.

Gemini’s evolution suggests the next AI battle will take place not only inside chat windows, but across the daily routines those windows are supposed to simplify.

Primary source: Google — The next evolution of the Gemini app.