AI & Models
Gemini Spark brings agentic AI to Google Workspace users
Google announced Gemini Spark, an agentic personal assistant integrating with Workspace, which it expects to make available to AI Ultra subscribers next week.
At its Google I/O developer conference on Tuesday, Google announced Gemini Spark, a new agentic personal assistant. Alphabet CEO Sundar Pichai described the product as a personal AI agent designed to help users navigate their digital lives by taking action on their behalf and under their direction. While traditional AI chatbots are limited to responding to direct text prompts, agentic AI represents a shift toward systems capable of taking autonomous actions. This new assistant is built using Gemini base models combined with an agentic harness—the underlying framework that enables autonomous capabilities—provided by Google Antigravity.
The launch places Google in competition with other AI developers that have introduced agentic products, most notably Anthropic with its Claude Cowork and OpenAI with its ChatGPT agent. However, Google’s primary differentiator is Spark’s integration with the Google Workspace suite. According to Josh Woodward, the VP of the Gemini App and AI Studio, this integration allows the assistant to pull facts directly from a user’s emails, docs, sheets, and slides to perform tasks. Woodward explained that if a user needs to send an email to their boss with a status update, Spark can pull all the facts from their emails, docs, sheets, and slides to write the draft. Additionally, Woodward noted that small businesses are using Spark to watch over their inboxes so they never miss a question from a customer.
To support these tasks, the assistant operates independently of the user’s local hardware. Pichai explained that “It runs on dedicated virtual machines on Google Cloud seamlessly, [so] you don’t need to keep your laptop open to make sure it’s running.” For mobile users, Google has introduced Android Halo, which serves as a system for tracking the agent’s progress on mobile devices.
To allow for broader utility, Spark can connect to external services using the Model Context Protocol (MCP), which is a standard for connecting AI to services. Google expects to roll out more connections for Spark in the months to come. Currently, Google expects to make Gemini Spark available next week to subscribers of Google AI Ultra, which is Google’s subscription service tier.
Why it matters
Gemini Spark differentiates itself from other agentic products by integrating directly with the Google Workspace suite. This allows the assistant to pull facts from emails, docs, sheets, and slides to perform tasks.