At the second edition of Google’s I/O Show for Android, Google announced major improvements coming to Gemini. The biggest of them involves how Gemini could bear the load of mundane or mindless tasks you do on your phone or other Android devices. This, as Google proposes, will be accomplished through Gemini Intelligence.

Although Google has been talking about these features for months, they were limited to basic chores, such as booking a cab or ordering your favorite meal on a food delivery app. With Gemini Intelligence, Google plans to take Gemini’s agentic capabilities to the next level, where it can also handle complex tasks, such as finding the most suitable guided tours that fit your specific requirements. And it won’t be limited to apps installed on your phone, because Google promises to bring agentic browsing to Chrome on Android as well.

Another application of Gemini Intelligence is the improved dictation mode, where Gemini would make sense of your ramblings and type out a well-organized piece of text.

More than 55% of the respondents in the survey say they aren’t impressed by or interested in Gemini Intelligence features. Another 25% are on the fence but aren’t fully sold on the idea either.

Perhaps that means Google needs to develop more convincing use cases for the next generation of Gemini. My colleague Joe Maring also believes Gemini Intelligence doesn’t feel like the most exciting evolution for Gemini as a whole, and he’s especially skeptical due to the hallucinations AI is generally prone to.

Additionally, Gemini Intelligence is also headed to your car with Magic Cue on Android Auto. While driving with Android Auto, Gemini could help you respond better by pulling up information, such as your upcoming dinner reservations, from other apps, like Google Calendar.

But to ensure that it works as intended, Google might need to work hard enough to iron out existing issues with Gemini, where it easily loses context or fails to provide acceptable responses, despite what Google claims.