Google Unveils Spark: $100/Month AI Agent to Manage Your Digital Life

Breaking: Google Launches Spark AI Agent

Google has formally introduced Spark, a personal AI assistant powered by its Gemini model. The cloud-based agent will operate around the clock, handling complex tasks across users' digital ecosystems.

Priced starting at $100 per month, Spark is now rolling out to trusted testers. A wider release to Google AI Ultra subscribers is expected next week.

How Spark Works

According to Google representatives, Spark connects to core Google services including Gmail, Docs, Sheets, and Slides. It can spawn sub-agents to tackle different aspects of a given assignment.

Example tasks include checking inboxes for client messages, auto-updating student study guides, or drafting emails based on data from multiple documents.

Third-Party Integration

Spark will support MCP connectors for services like Adobe, Asana, Box, Canva, Dropbox, HubSpot, and more. This allows access to external files and project statuses.

"Spark navigates across your digital life," a Google spokesperson said. "It works tirelessly in the cloud, never sleeping."

Pricing and Availability

The AI Ultra plan has been revamped: $100 per month for the base tier, up to $200 for the top tier—down from the previous $250 rate. AI Pro users are not included.

Spark communicates via text and email; users can include it in text chains or CC it on emails. A new Android Halo UI for live task updates is expected later this year.

Background

Previously known internally as "Remy," Spark was first hinted at in earlier reports. Google's I/O event on Tuesday confirmed the name and pricing.

The agent lives in the cloud rather than on local devices, meaning it cannot access local files directly. This design choice aims to protect sensitive documents.

What This Means

Spark positions Google as a direct competitor to other AI agents like Claude Cowork. Its always-on nature and deep integration with Google Workspace could reshape productivity.

However, the $100/month entry point may limit adoption to power users and businesses. The success of Spark will depend on how well it handles permissions and approvals—details still being finalized.