Google Introduces Gemini Spark AI Agent for US Subscribers

Google has announced the upcoming launch of Gemini Spark, a new personal AI agent designed to perform proactive tasks across a user's digital environment [1]. The tool, which operates in the cloud to remain active even when devices are locked or closed, will be available to Google AI Ultra subscribers in the United States starting next week [2].

Key takeaways

• Gemini Spark is a cloud-based agent powered by the Gemini 3.5 Flash model that continues to run in the background after devices are locked [1, 2].
• The agent integrates with Google Workspace tools, including Gmail, Docs, and Slides, to manage tasks like organizing statements or summarizing meeting notes [1, 2].
• For high-stakes actions, such as spending money or sending emails, the system is designed to request user permission before proceeding [2, 3].
• The Google AI Ultra subscription plan is priced at US$100 per month [1].
• Initial access begins with trusted testers this week, followed by a broader Beta rollout for US subscribers next week [1, 2].

Capabilities and Agentic Functionality

Gemini Spark represents a shift toward what Google CEO Sundar Pichai describes as the "agentic Gemini era," where AI moves beyond simple chatbot interactions to perform work on behalf of the user [3]. By utilizing the Gemini 3.5 Flash model, the agent can be taught specific skills, such as monitoring an inbox for school-related updates and consolidating those details into a summary for a partner [1]. It is also capable of setting recurring triggers, such as identifying hidden subscription fees within monthly credit card statements [2].

The agent is designed to handle complex workflows, including synthesizing meeting notes from emails and chats to generate polished Google Docs [2]. Beyond internal Google tools, the system supports Model Context Protocol (MCP) integrations, with initial connections to services like Canva, OpenTable, and Instacart launching immediately [2]. Google plans to expand these capabilities in the coming weeks to include features such as operating a local browser, creating custom sub-agents, and enabling users to communicate with Spark via email or text [2].

Why it matters

The introduction of Gemini Spark highlights Google’s strategic focus on proactive, "agentic" AI that operates autonomously to save users time on routine digital tasks [3]. While the technology is designed to be helpful, Google has acknowledged the potential for user anxiety regarding the power of such agents, noting that the system is built to require explicit confirmation for sensitive actions [2, 3]. As the company continues to invest heavily in its AI infrastructure, the rollout of Spark serves as a central component of its effort to transform the Gemini app into a universal, 24/7 assistant [2, 3]. Future updates are expected to bring Spark to the Gemini macOS desktop app, further extending its ability to automate local file tasks and desktop workflows [2].