Google Reveals New AI Agents With Autonomous Task Capabilities

Google unveiled AI-powered information agents at Google I/O 2026, its annual developer conference, that monitor topics continuously in the background and push relevant updates to users without requiring a query. The feature begins rolling out this summer to AI Pro and Ultra subscribers in the United States, with expansion to additional markets planned later.

The shift is notable because Google has operated on a single-session model for most of its history: a user submits a query, Google returns results, and the session ends. These agents change that flow. Users configure a topic once, and the agent runs persistently, synthesizing updates from multiple sources and flagging developments it determines are worth surfacing.

Search Features Now Include Proactive AI Responses

Users access the feature through AI Mode in Google Search. A natural-language prompt, such as "keep me updated on nearby movie tickets for The Mandalorian and Grogu," activates a background agent that monitors that topic continuously. Active tracked subjects appear in a user's AI Mode history, where they can be refined, paused, or removed at any time.

The agents are built to do more than notify. Google says they synthesize information from multiple sources, explain why a development may be relevant, compare different perspectives, and surface actionable context where applicable. The intended output is something ready to act on, not just read.

Unlike Google Alerts, which has primarily relied on keyword-based email notifications since 2003, the new AI agents are designed to interpret context, evaluate significance, and deliver structured insights instead of lists of links.

Stocks, Flights, Listings: What the Agents Are Built to Watch

Google has confirmed the agents can monitor:

• Stock market activity and specific companies
• Flight prices for upcoming travel
• Sports results and live events
• Breaking news and policy developments
• Housing and job market trends
• Weather and traffic conditions
• Entertainment listings, including movie ticket availability

The breadth of use cases positions these agents as a general-purpose background intelligence layer, not a tool built for one category of user.

Moreover, Google has not detailed what editorial or technical safeguards govern how these agents evaluate sources, determine significance, or handle conflicting information. In a system where no user query triggers a fresh evaluation, the reliability of what gets surfaced matters considerably more than in a standard search session.

Behind a Paywall at Launch

Access is restricted at launch. Only subscribers on Google's paid AI Pro and Ultra tiers will receive the feature initially, and only in the United States. Timing for additional markets has not been specified.

The restriction means the feature will not reach the vast majority of Google's users in the near term. Google handles billions of searches daily, but AI Mode and the tiers required to access these agents represent a narrow, paying segment of that base. Whether the company plans to extend access to free users, or on what timeline, has not been disclosed.

Read More

• OpenAI Drops GPT-5.5 — The Most Capable Model in the Room Is Live Now
• OpenAI Introduces Workspace Agents in ChatGPT — Build AI Automations Without Writing a Single Line of Code
• OpenAI and Microsoft End Exclusivity — What Their Amended Deal Means for the Future of AI
• OpenAI Launches ChatGPT Images 2.0 — The Image Model That Thinks Before It Draws
• Google Gemini Now Generates Files — PDF, Word, Excel and More, Directly From Your Chat