The assistant does everything by itself: Google opens the era of digital agents

The integration of new agent capabilities aims to redefine interaction with traditional platforms, starting with YouTube. The new Ask YouTube feature allows you to aggregate different video content into interactive answers, facilitating direct access to the most relevant segments. On the productivity front, the introduction of voice tools in Gmail, Keep and Docs allows documents to be managed and edited via verbal commands. The Workspace ecosystem is also enriched with Google Pics, an application for image generation and retouching based on the Nano Banana model, designed to allow selective manipulation of graphic elements. The hardware expansion instead includes the autumn debut of smart glasses equipped with Gemini, designed to offer localized and hands-free voice assistance.

The architecture that supports this transition is divided into multiple hardware and software levels. Among the new models presented, Gemini 3.5 Flash stands out, optimized to guarantee execution capacity with quadrupled speed compared to previous standards, supported by Gemini Omni, focused on multimodal generation starting from video input. For the software development space, the release of Google Antigravity 2.0 introduces a standalone desktop application geared towards creating autonomous agents. In terms of personal services, the ecosystem introduces Gemini Spark, a continuously active assistant, integrated with daily summary functions and context agents applied to Search, complete with an automated universal shopping cart system.

The parallel expansion of functionality raises the need to implement security and content verification protocols. To this end, Google is planning to extend the SynthID and C2PA tracking systems into Search, the Gemini app, and the Chrome browser, with the goal of making it easier to identify AI-generated materials. On the infrastructural protection and scientific research front, the deployment of the Code Mender agent developed by DeepMind for system monitoring is expected, together with the launch of the dedicated Gemini for Science platform.