Google's highly anticipated AI agent, Gemini Spark, is now being made available to subscribers of Google AI Pro in the United States. The rollout follows its initial launch in May, which was exclusively for AI Ultra subscribers. According to Google, the feature will soon expand to AI Pro members in other countries, though no specific timeline has been provided.

Gemini Spark represents a significant leap in Google's AI offerings, functioning as a personal AI agent designed to automate complex, multi-step tasks. Users can create and manage tasks that leverage a variety of connected services, including Google Workspace (Gmail, Calendar, Docs, Sheets, Slides, Drive, Keep, and Tasks), Google Search, Personal Intelligence, a remote web browser, a remote computer with code execution capabilities, chat functionality, and Canvas. This integration allows Spark to perform workflows that previously required manual intervention across multiple apps.

The introduction of Spark brings a new conceptual framework to Gemini: Tasks, Schedules, and Skills. A Task is the high-level goal a user wants to accomplish, such as planning a business trip. A Schedule allows users to set automated triggers, like conditions (e.g., flight delay) or time-based events, that initiate tasks. A Skill is a reusable set of instructions with additional context, enabling users to combine multiple actions—for example, using a 'Travel Booking' skill and a 'Gmail Writing' skill simultaneously to rebook a hotel and send a confirmation email. Google has stated that users can have up to 15 tasks running concurrently, with usage subject to the same compute-based limits as the rest of the Gemini platform.

Workspace Integration Details

The Workspace functionality of Gemini Spark has received particularly large updates since the initial launch. The agent can now perform a wide variety of actions across Google's productivity suite, many of which require explicit user confirmation for edits to shared documents.

Calendar

• Manage your schedule: Check your calendar, RSVP to invitations, and schedule new events.
• Reschedule & clean up: Update meeting times, change locations, or cancel events.
• Smart scheduling: Check your calendar and suggest open slots that fit everyone's availability.

Gmail

• Search & summarize emails: Search across your email threads to find specific topics or summarize conversations.
• Draft & reply: Compose email drafts, forward emails, or reply to existing email threads.
• Use labels: Organize your inbox with labels.

Docs, Sheets, Slides, & Drive

Important: To edit shared documents, spreadsheets, and presentations, you'll need to review and confirm Gemini Spark's planned edits. This safeguard prevents unauthorized modifications to collaborative files.

• Search & manage your Google Drive files: Search for files, read file contents, rename or delete files, check metadata, and view recent documents.
• Write & edit documents: Create documents in Google Docs or edit existing ones—adding summaries, notes, or editing specific sections.
• Read comments: Read comments in Google documents, spreadsheets, and presentations.
• Create & edit spreadsheets: Build, format, and edit custom spreadsheets in Google Sheets with data, tables, and formulas.
• Generate & edit presentations: Create presentations in Google Slides from a prompt, and edit, add, and remove slides in a presentation.
• Add images to documents & presentations: Add an image to your prompt and ask Gemini to add it to a document or presentation.

Google Keep

• Find notes & lists: Search for and retrieve notes and lists across Google Keep.
• Create notes & lists: Create private lists and notes in Google Keep.

Google Tasks

Important: Gemini can perform bulk actions on private tasks in Google Tasks without your confirmation. Users are advised to review requests before submission or manage tasks directly in Google Tasks.

• Find reminders & tasks: Search for and retrieve reminders and tasks from Google Tasks.
• Manage reminders & tasks: Add, update, and delete reminders and tasks in Google Tasks.

Expanding Capabilities and Context

The rollout of Gemini Spark to AI Pro users underscores Google's strategy of tiered AI access. The AI Ultra subscription, introduced earlier this year, offered the most advanced features, while AI Pro serves as a mid-tier option for power users and professionals. By extending Spark to Pro subscribers, Google aims to test and refine the agent with a broader user base before a potential general availability release. The feature is also accessible via the Gemini macOS app, which allows users to initiate tasks—such as browsing files—from their phone, showcasing seamless cross-device integration.

Gemini Spark's architecture builds on Google's broader AI ecosystem. The use of Personal Intelligence draws from user data across Google services to personalize task execution. Meanwhile, the remote browser and code execution components enable Spark to interact with web pages and run scripts, expanding its utility beyond simple API calls. This positions Spark as a direct competitor to other AI agents like Microsoft Copilot and OpenAI's ChatGPT plugins, which also aim to automate workflows across applications. However, Google's deep integration with Workspace gives it a unique advantage in the productivity space.

Since its initial launch, Google has continuously expanded Spark's capabilities. The Workspace improvements are the most notable, but the agent has also gained the ability to handle more complex conditional logic in schedules. For example, a user can set a schedule that triggers when a flight is delayed, automatically notifying the user and proposing an updated itinerary. These refinements demonstrate Google's commitment to iterating quickly based on user feedback and usage patterns.

From a technical standpoint, Spark runs on the same compute infrastructure as other Gemini services, and usage is limited by a compute-based quota. This ensures that while tasks can be numerous, they do not overwhelm the system. Google has not disclosed the exact limits, but the allowance of 15 concurrent tasks suggests a generous allocation for most users. The agent's ability to combine multiple skills in a single task points toward a future where AI handles increasingly sophisticated routines, from travel planning to content creation to data analysis.

The phased rollout—first to AI Ultra, now to AI Pro in the US—suggests that Google is carefully monitoring performance and user satisfaction before wider expansion. The company has not announced plans for free tier access, but given the compute costs involved, Spark is likely to remain a premium feature. Nonetheless, its availability to Pro subscribers marks a significant step in democratizing advanced AI automation.

Source:9to5Google News