[DigitalToday reporter Jinju Hong (홍진주)] Google appears set to add a "Proactive Assistance" feature to Gemini that suggests needed information before users ask questions.

On April 28, local time, IT outlet PhoneArena reported the feature was spotted during analysis of Google app version "17.18.22.sa.arm64".

The core is that Gemini reads a user's current context and shows recommendations at the right moment. Google is said to have indicated the feature uses what is shown on the screen, notifications and information from apps the user has directly connected to provide recommendations at appropriate times. Users can turn the feature on with a large toggle button and choose which apps to pull information from.

An example is specific. If a friend asks by text message to send the address of an upcoming event, Gemini would pull the address from an email and suggest it as a "smart reply" for a text response. It is designed to surface needed information up front without the user having to search across apps.

There are 3 sources it can pull information from: what is currently visible on the smartphone screen, various notifications, and apps the user has connected. The connected apps can include not only basic apps such as contacts and messages but also Google Workspace apps including Gmail, Calendar, Google Docs, Google Drive and Google Keep. Google explained it would provide reminders, recommendations, insights and action suggestions, and that such information should arrive in a useful moment rather than when users go looking for it themselves.

Privacy safeguards were also presented. Google specified that data is processed on the device rather than in the cloud and that information is stored in a "private encrypted space" on the device. It also said Proactive Assistance does not use data for generative AI model training or human review.

The feature also overlaps with "Magic Cue," which Google previously mentioned for the Pixel 10 series. Magic Cue was introduced as a function that understands what the user is doing and presents needed information at the right time. At the time, it was suggested it could become a latest Pixel-only feature based on the performance of the Tensor G5 application processor, but this time a similar approach appears to have been confirmed at the Gemini level.

Google is showing a trend toward broadening the focus of smartphone AI competition from simple question-and-answer to proactive recommendations. If a structure that interprets real-time context by combining screen, notification and productivity app data is applied to actual products, digital assistants would move beyond tools that respond after a user request to a stage of suggesting actions first.

Still, what has been confirmed so far is limited to a settings menu under development and feature descriptions. The actual release timing, the range of supported devices and whether it will be provided by default in Gemini have not been disclosed. That leaves the question of whether Google will limit the feature to Pixel devices or expand it across Android as a point to watch.