The global search engine platform is now introducing its own agentic AI solution to the Gemini App this year. Just like the OpenAI’s Operator, it will be able to accept prompt, create a checklist of things that should be done and then next do them for you.
The Google CEO Sundar Pichai made this announcement and introduced the new ‘Agent Mode’ that will allow users to make Gemini AI to do numerous tasks at once including search for a home in a specific locality and within a particular budget range. This tool is able to search the web by itself and give the result. It will even allow the users to book them through Project Mariner.
During the I/O keynote, Pichai asked the Agent Mode to look for an apartment in the nearby locality. The Gemini foundation model-powered AI agent started Zillow, a house-searching service, searched the site to set filters, and then it used MCP to book a showing of the house for him; everything was done without any human intervention.
Agentic AI has been a perfect solution for the industry, with every known brand making significant investment to automate how people interact the computer. While to access thing still needs human interaction, such as typing or speaking, agentic AI will be able to learn about a user and do the basic function with a simple command.
The bot uses AI to crawl sites, documents, and files to take action on the people’s behalf across different devices, and doing thing such booking a tick to adding items in the cart.
Pichai mentioned that this new feature is presently experimental, which means it may make mistakes, so you will have to be careful while relying on it. Because the Agent Mode will be available in Gemini, Chrom, and Search, it will perform numerous functions at once. Pichai has not explained yet how many tasks the Agent Mode will be able to do. He further mentioned that this new mode will be available for subscribers only.