Google opens Gemini tasks to free users in six countries
Editorial Team
••📖6 min read
Google is removing the subscription requirement for Gemini’s task mode in six countries, while adding U.S. payment, calling and messaging features. The expansion broadens access to delegated work, but usage limits, permissions and website compatibility still govern what it can finish.
Gemini’s task mode is rolling out free in the U.S., Japan, South Korea, Canada, Australia and India.
New U.S. features add Google Pay checkout and task requests by text or email; purchases require approval.
Free access has compute-based limits, and Google still calls the task capability experimental.
Google is opening the task mode in Gemini, its consumer AI assistant, to users without a subscription in six countries from November 17. The feature lets people delegate multi-step work across apps and websites. Alongside that expansion, Google is adding U.S. capabilities for payments, business calls and requests sent by text or email.
Google says in its announcement that the free rollout, powered by its new Gemini 4 Hydrogen AI model, will reach hundreds of millions of people. That describes potential access, rather than measured adoption. The commercial opening is to make Gemini the place where more users manage activity across Google and third-party services, while retaining paid plans for heavier use.
Google’s Task-mode interface, shown in its announcement of free access in six countries. Image: Google. Source: Google.
A paid agent becomes a free task mode
Tasks grew out of Gemini Spark, the personal AI agent Google began rolling out to trusted testers on May 19, 2026, with a U.S. beta for adult Google AI Ultra subscribers to follow. Google’s release history also records a June expansion to its Mac app. The capability later required AI Pro and was folded into the main Gemini experience in October, according to reporting on the change. Users now select Task beside the prompt box.
The mode can organize an inbox, research a trip, monitor prices or create documents, spreadsheets and presentations. Google’s help documentation distinguishes a task—the goal—from a schedule, which triggers it at a time or after an event, and a skill, which supplies reusable instructions. These are tools for recurring work as well as one-off errands.
Microsoft is pitching Copilot Workspaces as a shared environment for people and AI agents. Team collaboration is already in early access, but personal Autopilot integration is still coming soon, with final agent pricing unannounced.
Editorial Team
Free access requires a personal Google account, an age of at least 18 and History turned on. Work and school accounts are excluded. Elsewhere, tasks require Google AI Pro or Ultra in supported markets; the European Economic Area, Nigeria, Switzerland and the United Kingdom remain excluded.
The free allowance is not unlimited. Google says usage depends on task volume and complexity, with higher subscriptions increasing compute-based limits. Its task help page does not specify a daily free-task quota. The separate ceiling of 15 tasks running simultaneously is a concurrency limit, not an allowance of 15 tasks per day.
Google’s setup example shows texting and email channels for Gemini tasks. The new channels are rolling out in the U.S. Image: Google. Source: Google.
New ways to call, pay and send requests
The November 17 release notes describe several U.S. additions. Their restrictions matter independently of the six-country free rollout.
Capability
What Google says it does
Rollout or control boundary
Google Pay checkout
Stages an itemized purchase within a user-set budget
English, U.S. adults with personal accounts; explicit one-tap and device approval before purchase
Text requests
Starts tasks and accepts follow-up instructions through Google Messages or iMessage
English, U.S.; verified U.S. number and eligible carrier required
English, U.S. adults with personal accounts; replies privately with a review-and-confirm link
Business calls
Calls U.S. businesses, including multiple shops, and returns a summary and transcript
Users can review and approve the plan; Gemini identifies itself as AI and businesses can opt out
Google’s announcement excludes calling in Indiana, Louisiana, Minnesota, Montana and Nebraska. The release notes say the caller will not share payment or sensitive account details over the phone.
Google says purchases use its Agent Payments Protocol, or AP2, with digital tokens on supported cards. Users retain their card rewards and purchase protections, and receipts and tracking information go to Google Wallet. American Express support is described as coming soon, alongside Mastercard and Visa. Integration of the Universal Commerce Protocol, another planned step in the checkout system, is still ahead.
Not every shopping task ends in an autonomous purchase: Google’s grocery example hands the user to Instacart to complete the order. Its launch examples of bill analysis, appointment booking and shopping are company anecdotes, also relayed in coverage of the announcement, rather than independent tests of typical savings or success.
Google’s product demonstration shows Gemini reporting calendar updates and a newly created Slides presentation. Image: Google. Source: Gemini.
App connections bring reach and permissions
Google’s pitch rests on its existing relationship with users whose email, calendars, passwords, files and payment cards already sit in its services. It says Gemini connects with more than 60 popular apps. Its product page lists Google services such as Gmail, Calendar and Drive alongside at least 50 third-party apps, including OpenTable and Instacart. Connections are off by default, users choose what to enable, and availability varies by location, language and device.
Connected apps and computer-based execution already have competitive precedent. OpenAI, the developer of ChatGPT, launched ChatGPT agent in July 2025 with website navigation, code execution, editable slides and spreadsheets, and connectors including Gmail and GitHub. It initially rolled out to Pro, Plus and Team subscribers and supported recurring tasks. That history establishes overlap in capabilities; it does not establish the products’ relative reliability or their current price differences.
For Gemini, the permissions extend beyond reading information. With Google Workspace connected, tasks can send email, change calendar events and create or edit files. Google requires review of planned edits to shared documents, spreadsheets and presentations. It separately warns that bulk changes to private reminders in Google Tasks—the reminder-management service—can occur without confirmation.
Google says remote task execution uses an account-dedicated, hardware-isolated virtual machine, with sandboxing, prompt-injection filters and Safe Browsing. But isolation does not mean information stays entirely within Google: its help page says Gemini can share names, contact details, files, preferences and sensitive information with websites while acting for a user.
The documentation describes safeguards that seek approval before sensitive disclosures and consequential actions, while warning that malicious instructions in websites, emails or documents can mislead an agent. Users can stop tasks, take over the browser, disconnect apps and delete saved remote data. Google explicitly says those safeguards cannot eliminate every risk.
The next test is completed work after supervision
Execution also depends on the environment. Google’s browser documentation says local Chrome browsing requires the device and browser to stay awake and running. A remote browser can keep working after the device closes, although sign-in or other steps may require the user to take over. Mobile tasks use the remote browser.
Google acknowledges that tasks can fail and that some websites hand control back for the final click. It says it is working with websites and security providers to accommodate personal assistants on a web where bots have often been blocked.
The announcement supplies no measured completion rate or typical time saved. Those are the missing results needed to judge the expansion: how often tasks finish correctly, how much intervention they require, and how much useful work fits within free usage limits. Removing the subscription requirement widens the audience able to try delegation; it leaves the balance between effort saved and effort spent checking unresolved.
Google is rolling out file references and sharing for Gemini skills as adult personal accounts lose the ability to use Gems. Automatic conversion preserves supported material, but file limits and missing tools can interrupt existing workflows.
Muddy Waters disclosed a short position in Nvidia while questioning a possible connection involving senior adviser Ajay K. Puri. Its own account leaves the key identity question unanswered, separating its new claims from an existing server-diversion prosecution.
Nvidia reported $115.8 billion in quarterly revenue in its first earnings release since co-founder Jensen Huang resigned. Strong demand gives interim CEO Colette Kress a financial cushion, while a campaign for Huang’s return collides with unresolved investigations and the board’s reported opposition.
OpenAI launched GPT-6.2 Astra and Sol for professional work, cutting Astra’s standard input and output token prices by 20%. The models offer cheaper paths through selected tasks, but research evaluations, reporting failures and unfinished independent safeguard testing complicate the case for broader autonomy.
Microsoft’s first live speech-to-speech model offers OpenAI API compatibility and planned enterprise integrations. Its published evaluations show a low-cost option, with several rivals completing more agentic tasks.
Canberra-based Trellis Data is offering its 220-billion-parameter Agentaus 1 model to business and government. Its pitch goes beyond keeping data in Australia, but performance, deployment costs and a planned capital raising remain key tests.
Perplexity is adding direct collaboration between frontier models to Computer, letting a lead agent coordinate up to 16 supporting agents. Company-reported benchmarks suggest mixed teams can outperform repeated copies of one model, but leave customer costs and production reliability unresolved.
Marsh reported 105 client cyber incidents in the last two weeks of October, against 48 in the final two weeks of May. The broker links the increase partly to AI, while warning that insurance panel access may not guarantee an available response team.
Meta’s Muse personal AI agent has launched in Australia, with some local shopping and booking connections to follow. Its browser and payment tools broaden its reach, while the company’s technical disclosures qualify its privacy promises.
Microsoft and Salesforce are extending their sales software alliance with Copilotforce, combining CRM workflows with Microsoft 365 context under existing permissions. The Copilot plugin enters public preview, while Teams remains in pilot; earlier Microsoft tools and a Claude beta raise the bar for proving its value.
Microsoft expands MAI Image 2.7 from an editing preview into image generation and API access, with cheaper Flash pricing and a Canvas workspace. Benchmark claims depend on the workload, while the MAI Labs interface carries restrictions on commercial use.
OpenAI is widening applications for its legal AI configuration after improving research scores and expanding its search index. Access still requires approval, and benchmark results show substantial limits alongside the gains.
Microsoft has made Word, Excel and PowerPoint creation generally available inside Copilot, combining editable files with workplace permissions and collaboration. The broader shift toward autonomous work remains staggered across previews, while its claimed cost advantage rests on company-funded testing.
Knit Health’s NEA-led Series A funds deployment of Med-BEAM for care navigation and hospital operations. Its early company benchmark beats three Claude models at predicting clinical actions, while better patient outcomes remain unproven by that test.
Science Corporation’s Series D follows its European retinal-implant launch. Clinical results support improved central vision, but U.S. approval, eligible patients and commercial terms remain consequential questions for the startup, reportedly valued at $3.5 billion.
Google will let Gemini users save conversations while opting out of generative AI training on future chats. The global rollout also extends storage of history-off and temporary chats from 72 hours to around two months; human review and longer retention can still apply.
Eligible Pixel owners can join a Play Store beta to generate home-screen widgets from descriptions, but device, language and regional restrictions still apply.
Microsoft’s new in-house coding models compete with outside suppliers inside GitHub Copilot. Company tests show lower task costs at promotional prices, while its continuing financial ties to OpenAI complicate the rivalry.
Valkryn disclosed a $15 million seed round and early military testing contracts. Its evaluator-focused Battle Lab faces existing commercial tools, public testing capabilities and funding constraints that software alone cannot resolve.