Conversation Coach: A Voice-enabled AI System that Helps Practice Difficult Workplace Conversations

2026-08-31

AI Summary

为解决经理与员工有效沟通培训成本高的问题，提出了一种基于语音的AI系统Conversation Coach，通过模拟不同员工类型和提供个性化反馈来帮助经理练习困难对话。

Abstract

Effective manager-employee communication is critical for retaining high performers and developing underperformers, yet training managers in these skills remains costly. Text-based chatbots offer a scalable approach but cannot provide realistic rehearsal: managers need to practice speaking aloud to build confidence before high-stakes conversations. In this paper, we propose Conversation Coach, a voice-first AI system that enables managers to rehearse difficult workplace conversations in a realistic spoken format. The system addresses three challenges: achieving low-latency interactions with strong language understanding, enabling adaptive conversations through configurable bot personalities that simulate different employee types, and generating personalized feedback on content and policy compliance. We compare an end-to-end speech-to-speech model with a cascaded approach combining automatic speech recognition, a large language model, and text-to-speech synthesis. The end-to-end approach achieves 3$\times$ lower median (P50) latency with native barge-in capability at an estimated 8$\times$ lower cost, while the cascaded approach offers superior reasoning essential for coaching quality. We deployed the cascaded architecture in production, where 40,000+ managers used it over six months, with adoption patterns indicating selective use for difficult conversations.

Problem

• effective communication
• manager-employee
• training costs
• realistic rehearsal
• confidence building

Innovation

• voice-first AI system
• low-latency interaction
• adaptive conversations
• personalized feedback
• cascaded architecture