An autonomous AI agent that Google offers as a feature inside the Gemini app. It connects to Gmail, Google Calendar, Google Drive and Google Sheets, and carries multi-step work — cleaning up the inbox, researching a trip, reconciling invoices — through to the end in the background after you give the instruction. Processing runs in the cloud, so tasks continue even when your device is powered off. Work you repeat can be saved as a “Skill” and run automatically through “Schedules” triggered by time or by a condition. Unlike most AI agents, it is not a standalone app: it is included as part of a Google AI Pro or Ultra subscription.
Key Features
- Tasks: Connects to Gmail, Calendar, Drive, Docs and Sheets and executes multi-step requests such as “sort through last week’s unread mail and summarize the key points” from start to finish. No need to go back and forth in chat one question at a time
- Schedules: Runs tasks automatically on a time trigger (“every morning at 7”) or when a condition is met. Set a routine task once and leave it alone afterwards
- Skills: Define and save the steps of work you do often, then simply call it up to run it again. You can, for example, build a reply Skill trained on the tone of your own email
- Cross-app intelligence: Connects information across the Google apps you have linked and proposes or performs the most relevant action. It is particularly good at matching calendar entries against the contents of your mail
- Custom connected apps (MCP support): Connect third-party or self-built apps via an MCP server URL and fold them into a Spark workflow, reaching beyond the Google ecosystem
- Confirmation gate: The design always asks the user for confirmation before any high-impact action. Execution is autonomous, but the final decision stays with a human
Pricing
There is no standalone pricing plan for Gemini Spark. It is provided as a feature included in a Google AI subscription.
| Plan | Monthly price (as listed in JPY) | Gemini Spark |
|---|---|---|
| Free | ￥0 | Not included |
| Google AI Plus | ￥725 | Not included |
| Google AI Pro | ￥2,900 | Included (subject to region and language) |
| Google AI Ultra | ￥14,500 / ￥32,000 (two usage tiers) | Included, with top-priority access |
It launched exclusively for Ultra subscribers and expanded to Pro subscribers in the US on July 16, 2026. Availability by country and language is still rolling out; at the time of writing, the EEA, the UK, Switzerland and Nigeria are outside the covered regions. The Japanese plan comparison page does not list Gemini Spark, so check Google’s official guidance for availability in Japan. Use is limited to people aged 18 and over.
Pricing is current as of August 2026. Please check the official site for the latest pricing.
Pros & Cons
✅ Pros
- Automates work across all of Google Workspace — Gmail, Calendar, Drive, Sheets. The more your work already centers on Google, the greater the payoff
- Because it runs in the cloud, tasks progress even with your device switched off. You can hand over research at night and pick up the result in the morning
- Repeated work can be saved as a Skill, so the effort of instructing it drops to almost nothing the second time around
- The design inserts a confirmation step before important actions, which keeps fully automatic behavior in check
⚠️ Cons
- You have to grant broad access to mail, documents and photos, and privacy concerns have been raised repeatedly
- Available countries and languages are limited; the rollout began in the US and is still expanding in stages
- There is no free tier. A Google AI Pro or Ultra subscription is required, along with an age limit of 18 and over
- Integration outside the Google ecosystem depends on MCP, which is not as effortless as the first-party connections
Comparison with Similar Services
| Criteria | Gemini Spark | ChatGPT (agent features) | Claude (Cowork) | Microsoft Copilot |
|---|---|---|---|---|
| Provider | OpenAI | Anthropic | Microsoft | |
| Main ground | Google Workspace overall | Browser operation, general tasks | File handling, document creation | Microsoft 365 apps |
| First-party integrations | Gmail, Calendar, Drive, Sheets | Limited (via connectors) | Mainly local files | Word, Excel, Outlook |
| Background execution | Continues in the cloud | Tied to the session while running | Runs in the desktop app | Varies by service |
| Sold standalone | No (bundled with Pro / Ultra) | Offered by plan | Offered by plan | Offered by plan |
Who Is It For
- People whose work and personal life both revolve around Gmail and Google Calendar, and who lose time to sorting mail and juggling schedules
- People who want to automate fixed weekly or daily work such as drafting reports or clearing out the inbox
- People who want to hand over the whole chain of researching, summarizing and following up — trip planning, invoice reconciliation and the like
- People already subscribed to Google AI Pro or Ultra who want to try agent features at no additional cost
Summary
Gemini Spark is an autonomous agent running inside the Gemini app that takes a direct run at automating the everyday work of Google Workspace users. The combination of cloud execution and saved Skills makes it easy to build the kind of automation you can set up once and stop thinking about. On the other hand, it is worth understanding up front that it assumes broad access to your mail and documents, that availability is regionally limited, and that there is no free option. If you already subscribe to Google AI Pro or Ultra, the realistic path is to start with low-impact routine work, watch how the confirmation gate behaves, and widen what you delegate from there.