Breaking News
Google is testing a new capability for its Gemini assistant on Android that could allow users to complete tasks inside apps without directly interacting with their phones. The feature, currently under development, is designed to let Gemini place orders, book rides, and carry out other app-based actions on a user’s behalf.
The upcoming functionality were discovered in a beta version of the Google app, indicating that the company is moving closer to transforming Gemini from a conversational assistant into a more action-oriented digital agent.
Gemini gains “screen automation” abilities
According to a report by 9to5Google, references to the feature were found in version 17.4 beta of the Google app through an APK teardown, a method used to examine app code for unreleased features. The capability is described internally as “screen automation” and appears under the label “Get tasks done with Gemini.”
The feature reportedly allows Gemini to interact with supported apps by navigating on-screen elements, selecting options, filling in forms, and confirming actions. This would enable the assistant to complete tasks such as booking cab rides or placing orders, rather than simply guiding users through steps.
Google has internally codenamed the feature “bonobo.” While the company has not confirmed a public rollout timeline, such beta references typically signal features being actively tested. Google has also indicated that screen automation will initially work with only “certain apps,” suggesting limited compatibility during early stages.
User control, warnings and privacy safeguards
Alongside the new automation tools, Google has included several cautionary notices in the beta. The company notes that Gemini may make mistakes and emphasizes that users remain responsible for any actions performed on their behalf. Users will be able to stop the process at any point and manually take control if needed.
Privacy disclosures indicate that screenshots of Gemini’s interactions with apps may be reviewed by trained reviewers if activity tracking is enabled. Google says this process helps improve its services. Users are advised not to share sensitive information such as passwords or payment details through Gemini and to avoid using screen automation for emergencies or high-risk tasks. The beta also references another experimental feature codenamed “wasabi,” linked to managing a user’s digital “Likeness.” This appears connected to 3D avatar technology already used in Google Meet on Android XR devices.
Together, these developments suggest Google is preparing Gemini for a more hands-on role across Android devices. While the features remain under testing, they point toward a future where Android assistants can actively perform tasks, not just respond to requests.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.



