How to get real-time help by going Live with Search (blog.google)

Google has launched Search Live in English across the U.S., rolling out a new real-time, multimodal search experience in the Google app (Android and iOS) — no Labs opt-in required. By tapping the Live icon (or selecting Live from Google Lens), users can engage in an interactive voice conversation in AI Mode while sharing their phone camera feed; Search sees what your camera sees, answers questions aloud in real time, and links out to relevant web resources. Camera sharing is enabled by default in Lens Live, enabling instant back-and-forth visual queries for travel tips, hobbies (e.g., identifying matcha tools), troubleshooting electronics, classroom experiments, or picking a board game. For the AI/ML community this is a notable production deployment of multimodal conversational search: it blends real-time vision-language understanding, speech interaction, and retrieval-augmented responses at mobile scale. Key technical implications include grounding visual context into conversational queries, latency and UX trade-offs for live video input, and integration of search retrieval with generative models to cite helpful links. The launch highlights progress and practical challenges in building safe, responsive multimodal agents — from visual recognition and reference resolution to handling follow-ups and privacy considerations — signaling a growing focus on real-time, camera-enabled AI assistants.