The Google Remy AI agent is reportedly being tested internally as part of the company’s broader effort to turn Gemini into a more action-oriented assistant. According to a report from Business Insider, Remy AI is designed to handle work and daily tasks on behalf of users through Gemini.
The report said Remy AI is currently available in a staff-only version of the Gemini app. Based on an internal document reviewed by Business Insider and The Information from two people familiar with the project, the tool is described as a “24/7 personal agent” capable of acting on behalf of users.
A spokesperson for Google declined to comment, and the report did not specify whether or when the feature could become publicly available.
Google Remy AI Agent Expands Gemini Beyond Chat
The Google Remy AI agent appears to be part of Google’s larger strategy to move Gemini beyond standard chatbot interactions.
Google already offers agent-related capabilities through features like Agent Mode, although availability depends on subscription tier and region. According to the report, Remy is designed to go further by integrating across Google services, monitoring information most relevant to users, handling complex tasks, and learning user preferences over time.
The report did not identify which Google services are currently included in the internal employee testing phase.
Gemini Connected Apps Play a Major Role
Google’s support documentation outlines how Gemini currently connects with multiple apps and services to complete requests and deliver more relevant responses.
According to Google’s help documentation, Gemini Connected Apps currently include Google Workspace tools such as Gmail, Calendar, Docs, Drive, Keep, and Tasks. The connected ecosystem also includes services like GitHub, Spotify, YouTube Music, Google Photos, WhatsApp, Google Home, and Android utilities.
The report suggests Remy may build on this connected-app ecosystem to take more direct action for users inside these services.
User Control and Privacy Remain Key Questions
The report also highlights ongoing questions around user control, privacy, and autonomy.
Google’s Gemini Privacy Hub already provides controls that let users review and delete Gemini Apps Activity, adjust auto-delete settings, and manage whether data is used to improve Google AI systems. Users can also control app access permissions and manage information saved for personalization.
Google’s existing Gemini documentation already describes actions with different levels of user impact, including retrieving Workspace information, creating calendar events, sending messages, opening apps, and controlling smart-home functions.
According to Google Research, AI agents should have clearly defined human controllers, limited permissions, observable actions, and structured planning systems.
Meanwhile, Google Cloud guidance says AI agent activities should remain transparent and auditable through logging and clearly defined actions. The company also emphasizes limiting agent permissions based on purpose and user risk tolerance through the least-privilege principle.
Technical Details Around Remy Remain Unclear
Several important details about the Google Remy AI agent are still unknown.
The report did not explain the technical architecture behind Remy, which Gemini model powers it, or how much autonomy is currently being tested. It also remains unclear whether Remy can complete actions independently without user confirmation.
Because of those unanswered questions, the approval process and logging system for completed tasks remain uncertain.
The internal document reportedly describes Remy as a “dog-fooding” project, a term commonly used in the tech industry when employees test products internally before a broader rollout.
Remy AI Draws Comparisons to Emerging AI Agents
The report compared Remy’s concept to OpenClaw Github, an AI agent that gained attention earlier this year for autonomously replying to messages, conducting research, and taking actions on behalf of users.
According to the report, Sam Altman said in February that OpenAI was hiring OpenClaw’s creator.
The article also noted that Demis Hassabis has previously discussed the long-term goal of building a digital assistant, though Google has not confirmed whether Remy will eventually launch as a public Gemini feature.
Don’t miss out on our latest news—follow us for the latest AI news, breakthroughs, and insights that matter.