[Digital Today reporter Jinju Hong (홍진주)] Google’s AI chatbot Gemini is evolving beyond answering questions into a tool that automates repetitive everyday tasks such as schedule management, news searches and email organisation. It can check needed information at set times, store frequently used work guidelines, or connect multiple apps to hand off tasks.

On Oct. 13 (local time), IT outlet Engadget reported that Gemini can automate a significant part of daily life and work through features such as scheduled tasks, external service connections, skills and Gemini Spark.

The first is checking schedules and calculating travel times. Users can ask Gemini to check their calendar every morning and estimate travel time between appointments. For example, it can be set to check that day’s schedule at 7:30 a.m. and use map information to calculate total travel and driving time. Users need to connect required apps such as Calendar and email to Gemini.

Travel time estimates can differ from actual traffic conditions. Gemini makes estimates based on information it has obtained when generating responses, so real-time traffic conditions may not be reflected. It is recommended to check routes and travel time again in a map app before departure.

The second is regularly checking the latest news in areas of interest. Using Gemini’s scheduled tasks, users can set it to search for the latest information on a specific topic at set times and summarise key points. A representative example is asking it to search for Apple-related news every day at 9 a.m. Users can receive notifications when results are ready, reducing the need to search manually each time.

It is also possible to receive daily and weekly summaries regularly. However, the number of scheduled tasks that can run at the same time is limited to 10.

The third is recommending weekend plans and nearby places. Users can set Gemini to suggest weekend plans every Friday or Saturday. It can recommend places to visit such as restaurants and cafes, and location access permission can be used to find nearby places.

One point to note is that scheduled tasks do not update location information in real time to match a user’s movement. They can operate based on location information obtained when the task is first set, so if a user has moved to a different area, recommendations may not match the actual location.

The fourth is saving repetitive work instructions as skills. Google is reorganising its existing Gems into a skill-centred system to store instructions and preferences that can be used repeatedly across multiple conversations. This is useful when applying brand guidelines to a weekly newsletter or maintaining a specific writing style and format.

Users can save needed instructions as a skill and call it in a conversation using the format '/skill name'. Google said it plans to introduce a method to call skills using the '@' symbol instead of a slash. It can be used for repetitive document writing and editing because users do not need to enter the same conditions each time.

The fifth is ongoing work automation using Gemini Spark. Provided to Google AI Pro and Ultra subscribers, Gemini Spark is an always-on agent that can keep performing tasks even when a phone or laptop is turned off. It can be assigned tasks such as reviewing a week’s worth of emails and summarising important content, creating to-do lists, and adding appointments to a calendar.

Spark integrates with Gmail, Google Calendar and Google Drive. However, connections with those services are disabled by default, so users must enable them directly in the settings menu. Google explained that Spark is designed to seek user confirmation before executing important tasks. This is intended to prevent actions such as deleting emails or files during the automation process.

Gemini’s automation functions are expanding beyond reducing simple repetitive tasks such as schedule checks and information searches to connecting work across multiple services. However, app connections and permission settings are required by function, and some features may vary by region. Users need to check the supported scope and execution conditions before using them.