Gemini will automate file editing and meeting tracking on Mac

It has been revealed that Google has prepared more advanced artificial intelligence features on the macOS side for the Gemini application. The new system the company is working on will allow Gemini-supported agents to control the user computer, edit files, organize folders and carry out operations in connection with Google Workspace applications. The emerging details show that Google specifically aims to develop an experience similar to Anthropic’s Claude Cowork solution.

As of today, Gemini’s macOS application offers a very limited structure. The app works with a native chat interface that basically resembles the web version. In addition, there is also a keyboard shortcut that allows users to quickly access Gemini throughout the system. This interface, opened with the alt and space keys, allows the active window to be shared with Gemini, making it possible for artificial intelligence to analyze the content on the screen. However, the current version does not include direct computer control or automation capabilities.

New details revealed through APK reviews show that Google wants to take Gemini to a more assertive point among desktop productivity tools. Sample commands seen in the codes reveal that Gemini agents can scan files, analyze and edit contents, and complete some tasks automatically.

Google Workspace integration attracts attention

Among the sample prompts that emerged, file editing and Google Workspace integration particularly stand out. It is stated that Gemini agents can scan documents in local folders and automatically transfer the data to Google Sheets tables. In addition, it will be possible to group and archive irregular files in the desktop or downloads folder according to their types.

Another notable feature is the batch file renaming system. It is stated that Gemini can reorganize hundreds of mixed files under more readable folder structures by reading file metadata. It is evaluated that such automations can provide serious time savings, especially for users working with intense document flow.

The system Google is working on doesn’t seem to be limited to local files only. One example prompt involves taking transcripts or notes from the last Google Meet meeting and creating a summary follow-up email. Thus, it appears that Gemini can automate workflows by connecting services such as Docs, Meet and Gmail.

In addition to all these, it is stated that “Screen Access” and accessibility permissions on macOS will be used in order for the system to work. In other words, Gemini will be able to see the user’s screen and have some control over the mouse and keyboard. This approach has recently become widespread in artificial intelligence-based agent systems. Anthropic’s Claude Cowork solution can similarly perform operations directly on the computer.

Google’s work in this area is not entirely new. The company gave the first signals for agent-based computer use with the Gemini 2.5 Computer Use preview introduced last year. However, these features have not been made available to a wide range of users to date. On the Android side, the number of devices on which Gemini can perform in-app tasks is quite limited. For now, only certain models can automate simple operations.

Despite this, the new system prepared for macOS is expected to offer a more comprehensive experience. Particularly for organizations using Google Workspace, Gemini’s ability to automate processes such as file management, meeting tracking and document organization may be noteworthy. In addition, artificial intelligence gaining direct computer control may bring about new discussions on security and privacy. It is currently unknown when Google will make these features available.