After its initial launch in May with the premium AI Ultra tier, Google's personal AI agent, Gemini Spark, is now rolling out to Google AI Pro subscribers in the United States. The company has confirmed that the feature will 'soon' be available to AI Pro members in other countries as well, marking a significant expansion of its AI agent capabilities.

Gemini Spark is positioned as Google's answer to automating multi-step workflows that previously required manual intervention across various apps and services. On the web version of Gemini, users will notice a new tab at the top of the navigation drawer labeled 'Spark.' On Android and iOS, it appears as a dedicated page within the Gemini app interface.

What is Gemini Spark?

Unlike simple chatbots or single-task AI assistants, Gemini Spark functions as an agent that can execute sequences of actions across connected services. It leverages what Google calls 'Connected Apps,' including Workspace (Gmail, Calendar, Docs, Sheets, Slides, Keep, Tasks), Google Search, Personal Intelligence (context about the user), a remote web browser for online tasks, a computer with code execution capabilities, chat threads, and Canvas. These connections allow Spark to break down high-level goals into individual steps and carry them out without constant user oversight.

For example, a user could ask Spark to 'Plan and manage my business trip to London,' and the agent would search for flights, check calendar availability, draft emails to attendees, and even propose itinerary updates if a flight is delayed—all autonomously. The system introduces three core concepts: Tasks, Schedules, and Skills.

Tasks

A Task is the overarching objective the user wants accomplished. It acts as the container for the entire workflow. Examples include 'Plan and manage my business trip to London' or 'Research competitor products and create a comparison spreadsheet.'

Schedules

Schedules are automated triggers that launch tasks based on time or conditions. For instance, a user can set a schedule such as 'When my flight is delayed (condition), notify me and propose an update to my itinerary (task).' This brings a proactive element to the assistant, allowing it to anticipate needs.

Skills

Skills are reusable sets of instructions with additional context that can be combined. A user might create a 'Travel Booking' skill that knows preferred airlines and hotel chains, and a 'Gmail Writing' skill that drafts messages in a specific tone. These skills can be used simultaneously—for example, to rebook a room and send a confirmation email to the travel team.

Usage Limits and Performance

Google has implemented compute-based usage limits for Gemini Spark, similar to those already in place for other Gemini features. Users can have up to 15 tasks running concurrently. The exact quotas depend on the subscription tier, with AI Pro users receiving a generous allowance that is likely to be expanded as the service matures.

Workspace Integrations

Since the initial launch, Google has significantly expanded Spark’s capabilities within Workspace, making it one of the most deeply integrated AI agents available. The following are the key integrations:

Calendar

Spark can manage schedules directly: checking calendar availability, RSVPing to invitations, scheduling new events, rescheduling meetings, changing locations, and canceling events. It can also perform smart scheduling by scanning multiple attendees’ calendars to suggest open slots that work for everyone.

Gmail

Email management is fully supported. Spark can search across threads to find specific topics or summarize conversations. It can draft replies, compose new emails from scratch, forward messages, and organize the inbox using labels. Bulk operations are permitted but require user confirmation for sensitive actions.

Docs, Sheets, Slides, and Drive

Spark can search for files in Drive, read their contents, rename or delete them, and check metadata. It can create new documents in Docs, edit existing ones by adding summaries or notes, read comments, and modify specific sections. For spreadsheets, it can build, format, and edit custom spreadsheets with data, tables, and formulas. Presentations in Slides can be generated from a prompt, and slides can be added, edited, or removed. Images can be inserted into documents and presentations. Importantly, any edits to shared files require user review and confirmation before changes are applied.

Google Keep

Keep integration allows Spark to find existing notes and lists, as well as create new private notes and lists based on user requests.

Google Tasks

Spark can search for reminders and tasks, retrieve them, and add, update, or delete entries. Bulk actions on private tasks can be performed without additional confirmation, though users are advised to review requests before submitting them.

Cross-Device Continuity

The Gemini macOS app also supports Spark, allowing users to start tasks on their phone and continue them on a desktop, or vice versa. For example, a user might initiate a task to browse files on their phone while away from their computer, and the task will remain active and accessible across devices.

Google continues to refine Spark’s underlying models and connections, with the goal of making it an indispensable productivity tool for professionals who manage complex projects. The rollout to AI Pro subscribers signals a broader availability strategy, as the company aims to compete with similar agent-based offerings from other tech giants. With its deep integration into the Google ecosystem and the ability to handle multiple concurrent tasks, Gemini Spark represents a significant step forward in personal AI assistants.

Source:9to5Google News