en.Wedoany.com Reported - Google in the United States is developing a new AI personal agent for its Gemini large model, internally codenamed "Remy," which has now entered the employee internal testing phase. Multiple sources indicate that the product is positioned as "an all-day personal agent for work, school, and everyday life, powered by Gemini," a project that has sparked widespread attention in the U.S. tech community.

Internal documents describe Remy's goal as "upgrading the Gemini app into a true assistant capable of taking action on behalf of users, not just answering questions or generating content." The current Gemini app primarily provides users with information and content generation services through conversational interaction, whereas Remy is designed as an autonomous agent with "action capabilities"—the core difference between the two being that Remy no longer waits for every user command but proactively monitors matters of user concern, handles complex tasks, and learns user preferences over time.

Remy is deeply integrated into Google's suite of services, with direct access to core products such as Gmail, Chrome, and Calendar. It can automatically monitor key emails in the inbox, perform search operations in the browser, adjust schedules in the calendar, and optimize task handling methods through deep learning of user habits. Google has not yet commented on the Remy project, nor has it disclosed a specific timeline for the product's public launch.

The advancement of the Remy project is closely tied in timeline to the viral popularity of the open-source AI agent tool OpenClaw. OpenClaw rapidly gained fame earlier this year for its ability to autonomously perform tasks on personal computers, garnering over 100,000 GitHub stars within a week and being called "the next ChatGPT" by NVIDIA CEO Jensen Huang. The explosion of OpenClaw directly validated the massive consumer demand for autonomous AI agents, and Google's internal testing of Remy is widely seen within the industry as a strong response to this market signal.

The global AI agent market is currently forming a multi-polar competitive landscape. Besides Google, Anthropic has launched Claude Cowork, Meta is developing an AI agent codenamed "Hatch," and NVIDIA is also advancing its NemoClaw platform. Information technology research and consulting firm Gartner predicts that by 2026, 40% of global enterprise applications will be equipped with task-oriented AI agents, a proportion that was less than 5% in 2025. For enterprises, choosing which platform to build and govern AI agents upon is becoming a critical proposition determining the success or failure of digital transformation.

Google has not yet launched an autonomous AI agent product for general consumers. Previously introduced agent features like "Agent Mode" can handle multi-step tasks, but their functional scope and subscription coverage are limited. Remy's positioning is clearly higher than these tools, as it is planned as a persistent intelligence layer woven throughout the Google ecosystem. Android application decompilation analysis shows that Google app version 17.20 has added code clues regarding a "Gemini Agent," with its welcome message set to "What can I accomplish for you today," and the interface description further explaining that the agent can "use your connected apps and skills to complete tasks like communication, document sharing, and shopping on your behalf."

Google plans to hold its annual I/O developer conference later this month, where artificial intelligence and AI agents are widely expected to be one of the key showcase directions in the keynote speech. At that time, more functional details of Remy are expected to be publicly revealed for the first time.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com