Google I/O 2026: Gemini Spark announced as Google’s new AI agent for daily tasks

Google I/O 2026 is currently underway at the Shoreline Amphitheatre in Mountain View, California, and one of the bigger announcements from the event is a new AI agent called Gemini Spark. With this, Google is trying to push Gemini beyond simple chatbot-style interactions and move towards task automation across apps and services.  Also Read: What is Cursor? The AI coding tool SpaceX is spending $60 billion on

The company says Gemini Spark is designed to work like a personal AI agent that stays active in the background and handles tasks on behalf of users.  Also Read: Google Earth's flight simulator is now on browsers: Here's how to play

What Gemini Spark actually does 

Gemini Spark runs on Gemini 3.5 and uses something Google calls the Antigravity harness. The focus here is on handling ongoing tasks instead of just answering prompts one by one.  Also Read: That old phone in your drawer could soon power Google's data centres; Here's how

Google says the AI agent is deeply connected with Workspace apps like Gmail, Docs, Slides, and other Google services. Since it works through the cloud, Spark can continue running tasks even if your laptop is closed or your phone is locked. 

Gemini Spark AI agent can help with daily tasks. (Image: Google)

The company showed a few examples during the announcement. Users can ask Spark to regularly scan credit card statements and identify hidden subscription charges. It can also monitor emails from schools, pull out important deadlines, and send a daily summary. 

Google also says Spark can create workflows involving multiple apps. For example, it can collect meeting notes from emails and chats, turn them into a Google Doc, and draft a follow-up email around the same project. 

New app integrations and future plans 

Google confirmed that Gemini Spark will support more connected apps over time. New MCP integrations launching include services like Adobe, Canva, OpenTable, Instacart, Samsung, Vivo, Xiaomi, OPPO, Spotify, Capcut, and more. 

According to the company, Spark will later be able to use these integrations to complete actions directly for users. 

Gemini Spark partners (Image: Google)

Google also mentioned a few upcoming additions planned for the summer. These include texting and emailing Spark directly, creating smaller custom AI agents for specific tasks, and allowing Spark to operate the local browser. 

The company says users will still remain in control of what the AI can access. Spark is designed to ask for permission before actions involving payments or sending emails. 

Gemini app for macOS also getting updates 

Alongside Spark, Google also announced updates for the Gemini app on macOS. The desktop app will eventually support Gemini Spark later this summer. 

Google is also working on new voice-based interactions for the macOS app. During the demo, the company showed Gemini converting casual spoken thoughts into properly formatted text drafts directly on screen. 

Add Techlusive as a Preferred Source

The updated macOS Gemini app is available starting today, while Gemini Spark support and the new voice features are expected to arrive later this summer for users.