Google turns Gemini into an AI agent

Google is turning Gemini from an assistant that answers questions into a system that can work on its own, organize your information and complete tasks under your supervision. The announcement comes at Google I/O 2026, as the app surpasses 900 million monthly users across 230 countries and more than 70 languages, according to the company.

The update brings together new models, a redesigned interface and two agents with more active capabilities: Daily Brief, which prepares a personalized summary every morning, and Gemini Spark, which can run workflows in the background.

Gemini goes beyond answering questions

The difference is in the type of task. A chatbot waits for you to write and then responds. An agent can receive a broader instruction, consult multiple apps, organize information and move the work forward without you having to direct every step.

For example, you could ask Gemini Spark to review your meeting notes, find related information in emails and chats, prepare a document in Google Docs and draft the message to start the project. It can also analyze monthly card statements to detect new subscriptions or alert you to relevant changes.

Spark runs in the cloud, so it can keep working even if you close your laptop or lock your phone. It connects with tools such as Gmail, Docs and Slides, and will use MCP connections, a standard that lets agents interact with external services such as Canva, OpenTable and Instacart.

Its autonomy has limits. Google says Spark will operate under your instructions, you choose which apps to connect and it will ask for confirmation before sensitive actions, such as sending emails or spending money.

Gemini Spark will begin with a group of testers this week. Google plans to launch a beta next week for Google AI Ultra subscribers in the United States.

An automatic summary to start your day

Daily Brief gathers information from your connected apps and turns it into a short update for each morning. It can review urgent Gmail messages, check upcoming calendar events and add relevant follow-up details.

It does more than summarize. It also organizes and prioritizes information based on your goals, and can suggest next steps. Over time, your “like” or “dislike” reactions help adjust the type of content you receive.

The feature is beginning to roll out in the United States to Google AI Plus, Pro and Ultra users. To use it, you will need to turn it on and give it access to the relevant apps.

Gemini 3.5 Flash and videos with Gemini Omni

Google also introduced Gemini 3.5 Flash, the first model in a new generation designed to combine reasoning ability with fast responses. The company did not provide independent performance metrics in this announcement.

Gemini Omni is designed for video creation. It lets you combine text instructions, images and videos to edit scenes through conversation. Its announced features include:

• Applying cinematic-style zooms.
• Changing backgrounds from a written instruction.
• Using templates to create videos from a phone.
• Generating a personalized AI avatar with your appearance and voice.

Omni is beginning to roll out to Google AI Plus, Pro and Ultra subscribers worldwide. It will not initially be available to all free users.

A more visual interface and a Mac app

The app introduces a visual language called Neural Expressive, with fluid animations, more intense colors, new typography and haptic feedback on compatible devices. It also integrates Gemini Live directly into the main experience, so you can move from typing a question to having a voice conversation without switching modes.

Google says Gemini will start presenting some responses with images, interactive timelines, narrated videos and dynamic charts instead of showing only blocks of text. This redesign is already rolling out globally on the web, Android and iOS.

The macOS app is now available to download. During the summer, Google plans to add Spark so it can work with local files and automate desktop tasks. It is also preparing a voice feature that will use the content visible on screen to turn a spontaneous explanation into precise text, directly wherever the cursor is.

For you, the most important change is not the new design, but the level of access and autonomy Gemini is trying to gain. Its usefulness will depend on whether you are willing to connect your email, calendar and other apps, and on how well you control its permissions. What needs watching now is whether these agents complete tasks reliably and whether their confirmations are enough before they act on your behalf.