Shawn Wen
Shawn Wen is a core member of PolyAI and a researcher and practitioner in the field of conversational voice AI, focusing on the technical architecture and deployment challenges of enterprise-grade voice agents.
Timeline (last 90 days)
PolyAI's Winson Ng believes that voice conversation is much harder than most people imagine, and that the real difficulty lies not in reasoning but in adaptation
Compared to the text and digital world, voice scenarios introduce time as an entirely new dimension, and no model has yet been trained that truly understands the concept of time
Enterprise customers want agents to be as powerful as humans while also being as controllable as machines, which is a contradictory demand
Winson Ng believes that voice agents for consumer scenarios are already buildable today, but enterprise-level use cases have additional governance and compliance requirements that must be met
Winson Ng emphasizes that PolyAI's innovation lies not in the model itself but in the IO contract (how data is prepared and how input/output mechanisms are designed), and believes the underlying model is not that important
All Facts (5)
PolyAI's Winson Ng believes that voice conversation is much harder than most people imagine, and that the real difficulty lies not in reasoning but in adaptation
50%UnverifiedCompared to the text and digital world, voice scenarios introduce time as an entirely new dimension, and no model has yet been trained that truly understands the concept of time
50%UnverifiedEnterprise customers want agents to be as powerful as humans while also being as controllable as machines, which is a contradictory demand
50%UnverifiedWinson Ng believes that voice agents for consumer scenarios are already buildable today, but enterprise-level use cases have additional governance and compliance requirements that must be met
50%UnverifiedWinson Ng emphasizes that PolyAI's innovation lies not in the model itself but in the IO contract (how data is prepared and how input/output mechanisms are designed), and believes the underlying model is not that important
50%