Google tests Remy, a 24/7 AI agent that makes purchases and plans your life for you

Google is turning Gemini into a 24/7 AI agent that plans your life for you

* Google is reportedly developing a new agentic version of Gemini. * The new Gemini is called Remy and will act as a 24/7 AI agent across daily life. * Remy will run digital errands, monitor routines, and help use connected apps and services. Google has plans for its Gemini AI models beyond the role of chatbot and into something far more involved, according to a Business Insider report. The company is building an always-on AI agent version of Gemini designed to complete tasks for users even across third-party services with minimal intervention from the user. The AI agent project is called "Remy," and internal documents say the goal is to create "a true assistant that can take actions on your behalf" in every part of a user's life. The point is not to engage in conversation with the AI as with the current Gemini, but to have it do things for you in the background. Proactive Gemini The agent will be "Your 24/7 digital partner," able to communicate with others, send documents, make purchases, and proactively complete errands without waiting for instructions. There have been hints that this is coming in recent months. Gemini's Personal Intelligence features let the AI come up with answers using content from Gmail and other Google services, including making AI images of the user based on what's been uploaded to Google Photos. Remy appears to turn that information, called "personal context," into action. Rather than acting like a simple chat window, Remy's AI agents offer dedicated sections for ongoing tasks, scheduled actions, and jobs waiting for user input. Completed tasks can be pinned, renamed, and reopened later. Ultimately, this would make the AI agents something to continually engage with, not just when you feel up to having a conversation with an AI chatbot. Convenience with surveillance If you're wondering how much information Google's AI will be processing to do all this, the warnings attached to the agent are full of language explaining it's experimental and could "make mistakes and expose data unintentionally." Users are advised not to rely on it for professional tasks. Users will reportedly be able to manage or delete that information through settings, as well as disable connected apps and certain personalization features. But an AI that can genuinely organize parts of your life cannot operate in isolation. It has to know where you go, what you search for, who you talk to, what you buy, and how you spend your time. For some people, that level of integration will sound useful. For others, it edges uncomfortably close to the idea of outsourcing free will to software. But Google has plenty of rivals pushing their own AI capable of operating browsers and apps with minimal human oversight. But a lot of people have already filled Google's ecosystem with information. So an agent would weave in perfectly into services they may already depend on every day. The AI industry is moving away from systems that simply respond and toward ones that act continuously. Google appears determined to make Gemini one of the first major examples of that transition. Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

Google is working on a "24/7 personal agent" that sounds a lot like its answer to OpenClaw

Google is building a personal AI agent called Remy inside its Gemini app. Google doesn't have a fully autonomous AI agent yet, but it's working on one. According to Business Insider, which reviewed an internal document, the company is developing an AI agent codenamed Remy. It is currently being tested by employees inside a staff-only version of the Gemini app. Remy is described as a "24/7 personal agent for work, school, and daily life" that can take actions on your behalf, monitor things that matter to you, handle complex tasks proactively, and learn your preferences over time. Google has declined to comment right now, and no public launch timeline has been confirmed. OpenClaw went viral, and now Google wants a piece of that market OpenClaw is the free open-source AI agent that took the tech world by surprise earlier this year, racking up over 100,000 GitHub stars in under a week. It can respond to messages, conduct research, manage files, and automate tasks across your computer without any input from you. Recommended Videos It became so popular that Nvidia CEO Jensen Huang called it "definitely the next ChatGPT." In fact, the appetite for OpenClaw was strong enough to push secondhand MacBook prices up by 15% in China. OpenAI ultimately hired OpenClaw's creator. Now, Remy sounds like Google's attempt to build something with similar ambitions but as a polished, integrated product. Every major player is now in the AI agent race Google's Remy project is a confirmation that the AI agent space is now a full-on race. Anthropic launched Claude Cowork, which can handle your PC tasks without the complex setup OpenClaw requires. Meta acquired Manus AI and launched My Computer, a desktop agent that sorts files, runs apps, and sends emails on your behalf. Meanwhile, Nvidia is building NemoClaw, an open-source platform that lets businesses deploy autonomous AI agents regardless of hardware. Although OpenClaw itself has faced serious security scrutiny, with researchers warning of exposed admin panels, prompt injection risks, and credentials stored in plain text. We can expect Google's version to be a deeply integrated, privacy-conscious agent from a trusted platform, which might be what actually wins this market. Google's Remy is currently in a dogfooding phase, which is standard practice at tech companies where employees test products before they reach the public. The company will hold its Google I/O event later this month (May 19-20), where it is widely expected to showcase its next wave of AI products. Agents are likely to be a centerpiece at this event, and Remy may well make its first public appearance there if Google is ready to show its hand.

Google's New 24/7 AI Agent Can Make Purchases on Your Behalf

Currently being used and tested by Google employees, Business Insider reports that Remy will have power to do much more than what Gemini can currently. As evidenced inside of the latest Google app, a string of code mentions the ability for Remy to communicate with others, share documents, and even make purchases. In order to do this, Remy would utilize information from your chats, Connected Apps, Personal context, Personal Intelligence, Agent files, and location. No doubt, some users might be weary of an AI Agent being proactive on their behalf. Most users will want to ensure oversight over what Remy is actually doing. For now, we assume Google has built in tools to allow for this, but we won't know until the company makes any of this official. With Google I/O right around the corner, there is absolutely no doubt that the company will use that time to showcase the next wave of AI products/innovations. If you're a fan of AI, it should be a blast.

The rise of AI agents: Google's secret project Remy goes beyond chat

Synopsis: Google is internally testing Remy, a new always-on AI agent designed as a proactive 24/7 personal assistant. Unlike traditional chatbots, Remy can monitor information, learn user preferences, and handle complex tasks independently. Currently trialed by Google employees, it may debut at Google I/O 2026, signaling the rise of agentic AI. Google is making significant waves in artificial intelligence by internally evaluating an advanced autonomous system known as Remy. This new development, highlighted in recent tech coverage, underscores the rapid transition happening across the sector from basic conversational tools to sophisticated digital helpers that can take initiative and complete work independently. Positioned as a constant, round-the-clock companion, Remy is intended to assist people with professional duties, learning activities, and routine personal matters by going far beyond simple question-and-answer interactions. The system stands out because it can observe and track details that matter to each individual, gradually build an understanding of their habits and needs, and then carry out involved sequences of actions without needing frequent guidance. Early testing is happening exclusively among Google staff members using a restricted edition of the Gemini application. This internal trial phase, commonly called dogfooding in the tech world, helps the company refine the technology before considering any wider availability. Although no definite public rollout date has been announced, there is growing anticipation that aspects of this project might appear during the Google I/O 2026 event starting May 19. Experts note that Remy shares similarities with other pioneering projects like OpenClaw, which has demonstrated the ability to manage communications, gather information online, and move through different programs on its own. Such progress illustrates the mounting emphasis on what specialists term agentic AI intelligent platforms engineered to handle entire processes rather than merely provide responses. Google has already begun limited experiments with agent-like functions in its Gemini service in certain areas, allowing it to perform practical operations such as exploring housing options, adjusting search criteria, and arranging viewings. This momentum is visible throughout the industry. OpenAI appears focused on creating a specialized mobile device that leans heavily on background intelligence for smoother, more intuitive experiences that blend local device power with remote computing. At the same time, Anthropic is strengthening its Claude model with stronger abilities to manage coding projects and other hands-on responsibilities that require less constant human direction. As these innovations advance, the overall direction in technology points toward environments where people can entrust more responsibilities to smart systems. These agents would coordinate across platforms, organize information, and execute plans efficiently, potentially changing how individuals engage with digital tools on a daily basis. Remy exemplifies Google's effort to remain at the forefront of this transformation, promising more seamless and capable support for users in an increasingly complex world.

Google Tests New AI Agent To Take On OpenClaw With Advanced Task Automation: Report - Alphabet (NASDAQ:GO

Google is testing a new AI agent, internally named "Remy". This agent operates within a staff-only version of Google's Gemini app and can integrate with a range of Google's other services, reported Business Insider on Tuesday, citing an internal document. Remy is a 24/7 Gemini-powered AI agent that acts on users' behalf, handling tasks across work, school, and daily life, not just answering questions or generating content. According to the report, AI agent appears to be more advanced than existing tools and has similarities to OpenClaw, an AI agent that gained popularity earlier this year for its ability to perform tasks such as responding to messages or conducting research for users. While there is no confirmed timeline for Remy's public launch, the document refers to it as a "dogfooding" project, indicating that employees are testing the product before its official release. Google did not immediately respond to Benzinga's request for comments. AI Agents Surge, OpenClaw Faces Heat The development of Remy comes at a time when AI agents are gaining momentum in the tech industry. OpenClaw is an open-source AI agent platform that enables autonomous task execution and decision-making with minimal user input. However, OpenClaw faced challenges as Anthropic cut its access amid surging AI demand. Last month, Anthropic stopped covering Claude usage on some third-party tools under the standard $200 per month subscription plan as surging demand strains its compute capacity. Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Image via Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Big Tech's Personal AI Agents Are Coming for the to-Do List | PYMNTS.com

Google's version is codenamed Remy, Business Insider reported Tuesday (May 5). It runs inside the Gemini app and connects Google's broader suite of services, including search, email and calendar. It is described internally as a round-the-clock assistant for work, school and everyday life. To make room for it, Google shut down its previous AI agent experiment, called Mariner, Monday (May 4), folding that team's work into the new effort, The Decoder reported Wednesday (May 6). Google declined to comment to the publication. Meta's version is called Hatch, The Information reported Tuesday. The company has built practice environments where the assistant learns to navigate real consumer apps, including DoorDash, Etsy, Reddit and others before deployment. Hatch is scheduled for internal testing by the end of June. Meta is also building a shopping tool for Instagram that lets a user tap on a product in a video and complete a purchase without leaving the app, aimed squarely at TikTok Shop. Both companies are reacting to the same event. In January, an Austrian developer named Peter Steinberger released a free tool called OpenClaw that let people send a message on WhatsApp or Telegram and have software handle the rest, such as booking a meeting, drafting an email or running an errand online while they slept. It became one of the fastest-growing pieces of software in internet history, reaching 3.2 million users in weeks, The Next Web reported Sunday (May 3). It also had a ceiling. OpenClaw required users to install and run software on their own computer, following instructions that tripped up anyone without a technical background. One of the project's own team members warned that it was too dangerous for people who couldn't navigate a command line. Then in April, Anthropic, whose AI model powered much of OpenClaw, cut off flat-rate access and required users to pay separately for every interaction, Axios repoted April 6. The tool that had seemed like a breakthrough for everyday consumers stayed, in practice, a product for enthusiasts. Mark Zuckerberg said on Meta's most recent earnings call that he sees an opportunity to build a version of the OpenClaw experience that is more polished and easier to use, PYMNTS reported Tuesday. Remy and Hatch are that version. The core difference is where these products live. OpenClaw required installation on a personal computer. Remy lives inside an app Android users already have. Hatch will run inside Instagram, where more than 2 billion people spend time every day. There's no setup. The assistant arrives inside something people already open. That positioning matches where consumer behavior is heading. More than 60% of consumers in the United States used a dedicated AI platform in the past year, according to PYMNTS Intelligence. Neither Google nor Meta faces the cost problem that ended OpenClaw's cheap access. Both own the computing infrastructure that their assistants run on. When Anthropic raised the price of running OpenClaw, millions of users were left without an affordable option. Google and Meta are building for exactly that audience into the places those users already are. Google's annual developer conference, where the company is expected to say more about its agent plans, is scheduled for later this month.