Google’s AI enters its ‘agentic era’.

What could be the natural successor of chatbots, if not virtual agents? We have officially entered the ‘agentic’ era of the big tech companies specialising in artificial intelligence systems. Indeed, the buzzword of the industry giants, including OpenAI, Anthropic and Google, is agent - a new type of AI assistant, capable of ‘better understanding the world around us, thinking ahead and acting on our behalf, under supervision’.

Google, more precisely, is currently developing two projects: Project Astra and Project Mariner. The former is the prototype of the company's ‘universal agent’ AI; the latter, on the other hand, is a tool capable of taking control of the browser and using a Chrome extension to complete tasks - it is currently still in its early stages, with some tests being carried out with a group of ‘trusted’ testers.

The two products are not yet ready for eventual distribution to the public, the available demos are strictly controlled by the staff, leaving little room for the ‘mind’ of the artificial intelligence.