Google launched a major Gemini app redesign at its I/O conference, introducing the Neural Expressive design language with refined typography, smoother animations, and haptic feedback. The update brings Gemini Live directly into the main interface, letting users seamlessly switch between typing and voice conversations without hunting for separate modes.

The overhaul also adds regional dialect support and transforms how the AI communicates, moving away from text walls toward visual responses with imagery, graphics, and narrated videos. Daily Brief, a new AI agent, launches today for Google AI subscribers, automatically aggregating data from Gmail and Calendar to prioritize daily tasks and suggest next steps based on personal goals.

Gemini Omni debuts as a new model capable of generating video content from mixed media inputs, including customizable backgrounds, effects, and personalized avatars. The Gemini Spark agent, powered by Gemini 3.5, integrates deeply with Workspace apps and third-party services to automate recurring workflows like expense tracking and report generation. Spark enters testing this week before broader rollout to Ultra subscribers.

These updates position Google's assistant as more proactive and multimodal, potentially reshaping how users interact with AI across productivity and creative tasks. The emphasis on voice-to-draft precision and workflow automation suggests Google is targeting power users who need reliable, hands-free assistance throughout their workday.