Hello Gemini, could you please place a McDonald’s Big Mac order?
Recently, Gemini has seen a number of intriguing updates. For example, we just announced that it would soon have a floating button that lets you control it while using any smartphone app. Google’s AI may soon be able to utilize the apps that are loaded on your device, according to a breakdown of the most recent Google app beta.
Request that Gemini handle everything for you.
At the moment, you would manually launch Uber or any other ride-hailing service in your area, choose the drop-off location, and confirm your booking if you wanted to schedule a transport to your office.
But you won’t have to perform all of this manual labor in the near future.
According to reports, the Google app beta version 17.4 contains a new “Get tasks done with Gemini” feature that Google is developing. This feature’s strings reveal that it is codenamed “bonobo” and will only be accessible as an experimental feature at Google Labs at first. In essence, Gemini will be able to use screen automation to communicate with specific installed apps and do activities on your behalf.
For example, you can tell Gemini to order specific meals from a food delivery platform like Uber Eats or arrange a ride to a specific place. All of that without requiring you to manually launch applications and select various settings.
It makes natural that this feature will initially only be accessible for a small number of apps, primarily those where users are most likely to utilize it, as app developers frequently alter the user interface of their own applications. I think the tech giant will begin implementing this screen automation feature in its own apps before releasing it in third-party ones, even though the deconstruction doesn’t mention it.
However, it has the same basic disclaimer that applies to all AI tools: it is prone to errors. You will allegedly notice an extra warning for this specific function, though, noting that you are in charge of any activities Gemini conducts on your behalf in accordance with your instructions. You should thus keep an eye closely on this feature.
You can stop the agent and take over the task if you discover that it isn’t operating as intended.
Because the screen automation function is based on Android 16 QPR3, your device must have at least this version loaded in order to take advantage of all the features listed above. The massive Mountain View company is also putting privacy first, as evidenced by its recommendation that you not give Gemini any personal information, such as login credentials or payment information. In order to enhance its overall services, the business will let qualified reviewers look at the screenshots Gemini captures of the apps it interacts with if you have the Keep Activity option set.
When Google first released Gemini, it was limited to simple tasks like fixing language in drafts or solving math problems. However, it can now accomplish considerably more because to the inclusion of agentic skills, and it doesn’t need much manual interaction from you.
For example, Google Chrome just got the Gemini 3-powered Auto Browse feature. It simply offers the browser the ability to manage tedious tasks like automatically filling out forms, but it is currently only available on the Pro and Ultra tiers. Gemini on Android is getting a similar functionality.
The majority of manual tasks have already been replaced by AI chatbots. For example, creating a trip itinerary no longer requires you to manually navigate across several websites. Alternatively, you can ask any AI chatbot to take care of that.
With the ability to reserve rides soon and place meal orders simply by giving Gemini instructions; in the future, we won’t need to execute these tasks by hand. Depending on how much a person wants AI to control their smartphone, this may or may not be a good thing. For my part, I am thrilled about this shift.
