Exploring the Potential of Gemini's Scheduled Actions Feature

Major Insights from Testing

When I first encountered the idea of directing AI to complete tasks on a set schedule, I was naturally intrigued. Google's Gemini has introduced a 'scheduled actions' feature that attempts to fulfill this function. However, while the promise is there, the current scope of what it can achieve is somewhat restricted.

Upgrading to Gemini's Pro plan unlocked the scheduled actions capability for me. While the concept of AI-driven automation is appealing, particularly for routine tasks, this initial iteration left me wanting more in terms of sophistication and range.

Using Scheduled Actions

Setting up an action within Gemini is designed to be simple. You basically incorporate a time and frequency phrase within your directive, like saying, "Each day at 5 a.m., perform the following…" With this, Gemini will add the directive to its scheduler; however, there's a cap of 10 active schedules, which requires strategic selection of tasks.

I adapted an earlier weather-related prompt from my experiments with other platforms and sent it to Gemini. After a brief processing period, Gemini acknowledged the task setup was successful, providing users with a satisfactory outcome.

Adjusting Existing Scheduled Actions

Should there be a need to alter a set action, it's a straightforward procedure. Navigate to the settings located at the lower left corner, select 'Scheduled Actions,' and from there, you can enable, disable, or amend any action.

Choosing the pencil icon allows you to adjust specific settings of the action, offering a well-structured and intuitive editor interface.

Potential Uses for Automation

The range of tasks that could benefit from Gemini's scheduled automation is vast. Here are fifteen creative starting points to consider.

Current Shortcomings in Image and Notification Handling

Two significant challenges emerged during my usage: limitations in handling images and notifications. If the schedule includes an image generation task, Gemini merely delivers the image without additional context. For example, in a weather-related task, an image request resulted in a pictorial response featuring a banana artistically styled with weather elements, but it failed to provide the desired informative context.

In terms of notifications, using the desktop version offers minimal alerting with just a small blue dot to denote activity completion—contrast this with the more overt alerts on mobile devices. Thus, the accessibility of notifications can vary significantly depending on the platform used.

Despite linking it to Google Workspace with the goal of receiving email notifications, there was no follow-through from Gemini on this front; email alerts were never received despite assurance from the system.

Room for Improvement, Yet Optimistic

It seems somewhat surprising that, given Google's capabilities, especially with Chrome and Gmail, efficiently leveraging these platforms for enhanced Gemini notifications and sending prompts via email remains underdeveloped.

Effective notification systems could greatly enhance the usefulness of automated actions, offering updates on significant news or industry trends. Currently, the need to manually check for updates diminishes the appeal of automation.

There's hope that over time, Gemini's abilities could expand to include REST requests or API integrations, allowing it to trigger actual-world changes or responses to specific scenarios. For now, though, the feature remains limited to running pre-set prompts.

Have you had a chance to experiment with Gemini's scheduled actions or similar tools? Were your expectations met, or do the current constraints affect your willingness to rely on it? What improvements would be necessary for you to use such features more regularly? Feel free to share your experiences and suggestions in the discussion below.