Voice control in robotics has traditionally required intricate coordination between speech recognition, natural language processing, and response generation. This session demonstrates a project developed together with the Hong Kong Institute of Information Technology (HKIIT), showing how Agentic AI and speech-to-speech streaming can simplify voice control for humanoid robots, robot dogs, and drones. The presentation explores practical AWS-based architectures designed for efficient command processing, intent recognition, and robot control systems. Through live demonstrations, learn how to minimize latency while maintaining precision across diverse environments. Discover patterns for handling continuous speech input and coordinating responsive robot behavior in real-world scenarios.