Agora CEO Zhao Bin outlines five-level conversational capability framework at iRTE2026
Agora founder and CEO Zhao Bin presented a "five-level conversational capability" framework during a speech at the iRTE2026 Real-Time Intelligence Conference, laying out an evolutionary path for conversational AI from basic response generation to deeper understanding and relationship-based communication. According to the event remarks cited by Techub News, the framework breaks conversational ability into five levels, starting from expression based on the system’s own perspective and moving toward communication built on deep relationships. The stated goal is to enable AI systems to do more than interpret the immediate context of a conversation, extending that ability to long-term context so they can better understand a person’s real intent. Zhao also said that as AI shifts from single-turn question answering to sustained interaction, real-time interactive infrastructure is becoming an important technical layer linking models, agent runtime frameworks, and users. Agora said its conversational AI engine is built around full-duplex interaction to support stable and synchronized real-time conversations. The company added that it will continue improving conversational AI development tools and toolkits for developers building agents with complete logic and execution capabilities.








