Google Tests Remy AI Agent for Gemini as Focus Shifts to User Control

According to Business Insider, Google is testing Remy, a new AI personal agent for Gemini. This tool aims to execute tasks on behalf of users, streamlining both professional workflows and daily routines.

Currently, Remy is undergoing testing in an internal, staff-only version of the Gemini application. The report cites an internal document and interviews with two individuals familiar with the project. Internal materials describe Remy as a “24/7 personal agent,” positioning Gemini as an active assistant capable of acting on a user’s behalf.

Sources close to the project confirm that Google employees are actively testing Remy. A Google spokesperson declined to provide further comment. The report offers no details regarding a public release timeline or whether such a rollout is planned. It also omits information about which specific Google services are included in this internal employee trial.

Task-taking assistant

Remy represents Google’s broader strategy to expand Gemini beyond simple chat-based interactions. While Google already provides agent-related features, such as Agent Mode, availability varies depending on subscription levels and geographic regions.

The report characterizes Remy as a more sophisticated iteration, designed to integrate seamlessly with Google services. It monitors user-relevant data, handles complex tasks, and learns from user preferences to improve performance.

Gemini’s connected-app surface

Google’s Gemini support documentation outlines the current scope of connected services, enabling Gemini to interact with other platforms to fulfill user requests and deliver more relevant responses. Connected Apps include core Google Workspace services (Gmail, Calendar, Docs, Drive, Keep, and Tasks), as well as third-party integrations such as GitHub, Spotify, YouTube Music, Google Photos, WhatsApp, Google Home, and Android utilities.

Control questions

Google’s Gemini Privacy Hub provides context for operations involving connected apps, including Google’s own services and third-party platforms. Users can review and delete Gemini Apps Activity, adjust auto-delete settings, and control whether their data contributes to improving Google AI. The hub also allows users to manage access permissions for other apps and data, as well as information explicitly saved by Gemini.

Existing Gemini documentation covers actions with varying levels of user impact, such as retrieving information from Workspace apps, creating calendar events, sending messages, opening applications, and controlling device or smart-home functions.

Google Research emphasizes that AI agents should operate under well-defined human oversight, possess carefully limited powers, perform observable actions, and demonstrate planning capabilities.

Google Cloud has stated that agent activities must be transparent and auditable through logging and clear action characterization. Their guidance stresses limiting agent powers based on intended purpose and user risk tolerance, adhering to the principle of least privilege.

Remy’s reported ability to learn user preferences highlights the importance of memory controls. Google’s Privacy Hub allows users to manage information saved by Gemini and provides controls for personalization based on past conversations and Personal Intelligence.

The report does not disclose technical details regarding Remy’s architecture, the underlying model version, or the level of autonomy being tested. It also fails to clarify whether Remy can act independently without user confirmation. These unanswered questions leave uncertainty regarding how Remy handles approvals and logs completed actions.

The internal document describes Remy as a dog-fooding project, a term commonly used in the tech industry when employees test products before a broader release. The report compares Remy’s concept to OpenClaw, an AI agent that gained attention earlier this year for its ability to autonomously reply to messages, conduct research on behalf of users, and take independent actions.

According to the report, OpenAI CEO Sam Altman announced in February that OpenAI was hiring OpenClaw’s creator. Google DeepMind CEO Demis Hassabis has previously discussed the goal of building a digital assistant, but Google has not confirmed whether Remy will become a public feature of Gemini.

See also: Google made agentic AI governance a product. Enterprises still have to catch up.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.