Voice AI Leaders Say the Technology Still Awaits Its Breakthrough Moment

During a recent HumanX conference, PolyAI’s chief technology officer Shawn Wen noted that voice AI has progressed to full-duplex models that can speak while listening, but it still lacks the rapid reasoning needed for seamless dialogue. He argued that future success depends on making responses fast enough to feel natural and on building user confidence in AI-driven customer service. Otter’s chief marketing officer Alex Gay pointed out that accurate speaker identification, intent detection, and transcription are critical for turning meeting recordings into useful, automated workflows, and he described the company’s work on digital twins that must convey human-like emotion.

Both leaders warned that poor automatic speech recognition can undermine downstream actions and erode trust. They also called for clear disclosure when AI is present in calls or meetings, suggesting notifications to all participants. The consensus is that voice AI must improve both technical performance and transparency before achieving a mainstream breakthrough.