Lgenie powers AI-native assistive devices with Agora’s real-time conversational AI

AIoT / Conversational AI / Elderly Care / Smart Devices

Problem

Creating natural, reliable voice experiences for assistive devices in challenging network conditions

‍

Solution

Agora Conversational AI enabled low-latency, full-duplex conversations with reliable streaming, real-time speech recognition, and seamless interruption handling

Results

400 ms avg response latency enabling near real-time conversation

100% accuracy with higher stability and completion rates across conversations and IoT tasks

100% success rate in interruption handling

About Lgenie

Enabling intelligent, low-latency voice interaction for elderly care and smart devices

Lgenie is an AI Agent platform purpose-built for smart devices and real-world AIoT applications. The platform enables developers and hardware partners to build and deploy AI-powered interactive products across smart elderly care, AI education, smart home devices, AI toys, and intelligent mobility scenarios.

One flagship initiative is Lgenie’s AI-powered smart cane, designed to provide elderly users with more natural and reliable conversational AI interaction in real-world environments. By combining conversational AI, voice interaction, and real-time sensing capabilities, Lgenie helps users navigate daily life more safely, confidently, and independently.

To support these experiences, Lgenie required a real-time conversational AI infrastructure capable of delivering low-latency interaction, stable mobile connectivity, and reliable voice performance across outdoor and weak-network environments.

The Problem

Delivering reliable conversational AI where responsiveness and trust matter most

For Lgenie, building AI-powered assistive devices for elderly users required a much higher level of responsiveness and reliability than traditional consumer voice applications. Since devices operate in outdoor and mobile environments, delays, interruptions, or unstable voice performance could directly impact usability and user trust.

Key Challenges

Delivering low-latency, human-like AI conversations

Maintaining stable voice interaction across outdoor and mobile 4G environments

Improving interruption handling and voice clarity in noisy real-world scenarios

The Solution

A production-grade conversational AI infrastructure for AI-native devices

Lgenie integrated Agora Conversational AI as the real-time interaction backbone for its AI-powered smart devices. By combining Agora’s Conversational AI Engine, RTSA, and ASR capabilities, Lgenie built a low-latency, full-duplex conversational pipeline capable of supporting natural voice interaction with high reliability.

Agora’s infrastructure improved conversational responsiveness, interruption handling, and voice clarity across real-world environments while reducing deployment complexity for AIoT devices.

Agora Products

Agora Convo AI Device Kit Agora Convo AI Device Kit

Why Lgenie chose Agora:

Ultra-low-latency conversational AI enabled near real-time interaction

Stable real-time performance improved reliability across outdoor and weak-network environments

Real-time ASR and interruption handling improved conversational fluency and voice accuracy

“Real-time interaction for older users leaves no room for delay or failure. Agora provided us with the reliability and audio quality we needed to build a device people can genuinely trust in their daily lives.”

Margo Wang, Marketing Director, Lgenie

‍

Results and ROI

Transforming assistive hardware into intelligent conversational AI devices

With Agora powering its conversational AI infrastructure, Lgenie transformed its smart cane into a more responsive and reliable AI-powered assistive device that delivers natural, real-time interaction for elderly users.

The platform achieved near real-time conversational responsiveness, improved interruption handling, and attained greater conversational stability across real-world environments, helping to improve usability and user confidence.

Results after implementing Agora:

400 ms average response latency enabling near real-time conversation

100% accuracy with higher stability and completion rates across conversations and IoT tasks

100% success rate in interruption handling

‍

TEN

App Builder

フレキシブルクラスルーム