● Covered by 1 source · 1 reportLow impact1 neutral
Voice AI Lacks a "ChatGPT Moment" Despite Full-Duplex Models, Say Industry Execs
Key points
- Voice AI has full-duplex models but lacks fast reasoning.
- Models need to fetch answers quickly for natural conversation.
- Customer service AI agents must sound less robotic.
- ASR models often miss important keywords, affecting context.
Current State of Voice AI
The voice AI sector has seen significant investment and development, with new models and tools frequently released. These advancements include full-duplex models, which allow AI to speak and listen simultaneously. However, industry leaders indicate that these developments have not yet led to a transformative moment for voice AI, similar to the impact ChatGPT had on text-based AI.
Challenges in Voice AI Development
Shawn Wen, CTO of PolyAI, states that while full-duplex models are a milestone, the next challenge is to make AI reasoning very fast. This speed is crucial for models to retrieve answers quickly, making conversations feel natural and fluid. Without rapid reasoning, the conversational experience remains limited.
Another significant hurdle is the accuracy of Automatic Speech Recognition (ASR) models. Wen notes that ASR often misses important keywords, which hinders the AI's ability to capture the full context of a conversation. Alex Gay, CMO of Otter, concurs, highlighting the ongoing effort to improve transcription and understanding.
Impact on Customer Service and Meetings
For customer service applications, Wen emphasizes that AI agents should not sound robotic and must instill confidence in callers that their problems can be resolved. He believes that once voice quality is sufficient and customers engage for a few turns, they will build confidence and potentially prefer AI over human agents for problem-solving.
In meeting environments, Gay from Otter points out the importance of speaker identification, intent capture, and integrating organizational knowledge for automation. Otter is also developing digital twins for meetings, where the output voice must convey emotive expressions to foster debate and strategic discussions, replicating human interaction quality.
The Need for Deeper Understanding
Both executives agree that current voice AI models often struggle with understanding users or producing accurate transcripts and summaries. The inability of ASR models to consistently capture critical keywords directly impacts the AI's overall comprehension and the utility of tools like meeting notetakers. Continuous improvement in these areas is essential for voice AI to achieve broader adoption and impact.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
Industry executives believe voice AI has not yet reached a breakthrough comparable to ChatGPT, despite advancements like full-duplex models. The primary challenges remaining are achieving faster reasoning for natural conversations and improving accuracy in understanding and transcription. These improvements are necessary for voice AI to build user confidence and handle complex interactions effectively.