Shengwang CEO Zhao Bin proposes a five-level conversational capability framework, outlining the path of conversational AI toward AGI

On October 23, the iRTE2026 Real-Time Intelligence Conference was held in Beijing. Agora founder and CEO Zhao Bin delivered a keynote speech titled "Conversing with AGI: The Next Chapter of Real-Time Interaction" at the main forum, reviewing key changes over the past year in real-time interaction, conversational AI, and underlying infrastructure, and proposing a "five-level conversational capability" framework that outlines the evolution path of conversational AI from basic responses toward deep understanding and relationship-based communication.

Zhao Bin believes that as AI moves from answering questions to continuous exchange, understanding context, and executing tasks, the object of real-time interaction is expanding from person-to-person to person-to-Agent. The role of real-time interaction technology is extending accordingly: it must not only make information transmission more stable and smoother, but also provide real-time support for AI's perception, understanding, expression, and action.

From "one question, one answer" to deep relationship communication: the five-level evolution of conversational AGI

In Zhao Bin's view, the progress of conversational AI should not be measured only by response speed and language naturalness. More importantly, it is whether it can understand the intent, situation, and relationships of the conversation partner, whether it can express itself appropriately in complex scenarios, and whether it can gradually form long-term, in-depth understanding.

Based on this judgment, Zhao Bin proposed in his speech a "five-level conversational capability" framework, dividing conversational capability into five levels:

L1: Expression based on one's own perspective: mainly self-expression and passive continuation, achieving "one question, one answer."

L2: Expression based on interaction patterns: moving from passive response to proactive expression, making conversation more natural and coherent.

L3: Communication based on individual cognition: understanding an individual's identity, intent, and situation, achieving communication tailored to the individual.

L4: Communication based on group cognition: understanding social relationships and role differences, adapting to complex multi-person communication scenarios.

L5: Communication based on deep relationships: understanding long-term context and relationship evolution, forming deeper understanding and tacit thinking, and moving communication from understanding the present to understanding a person.

What these five levels show is not only the step-by-step enhancement of conversational technology, but also the change in the relationship between AI and people: from completing a round of Q&A, to understanding a person, to understanding relationships between people, and ultimately forming deep communication across time and situations.

Zhao Bin pointed out that conversational AI is developing from early passive-continuation interaction toward more proactive, more natural full-duplex conversation, and is beginning to try to understand people's personalities, situations, emotions, and relationships. Looking to the future, AI also needs to break through role division and communication order in complex social scenarios, and further understand the context accumulated by an individual over time. This is also an important development direction for conversational AGI: AI can not only understand a sentence, but also combine the full context to understand a person's true intent and respond appropriately.

Real-Time AI Infra: key infrastructure supporting conversational AGI

The continuous evolution of conversational capability also places higher demands on underlying technology.

Zhao Bin mentioned in his speech that over the past year, deeper adoption of WebRTC technology by companies such as OpenAI has driven broader attention to and application of real-time interaction technology in the conversational AI field. As AI moves from single-turn Q&A to continuous interaction, real-time interaction infrastructure is becoming important technical support connecting models, agent runtime frameworks, and users.

Compared with traditional AI applications, conversational scenarios have stricter requirements for latency, stability, and two-way interaction. AI needs to receive user input while understanding context and organizing responses, and when necessary respond to interruptions at any time while continuously calling tools and advancing tasks. Significant latency in any link may undermine the naturalness of the conversation and even affect whether the task can be completed.

To support these needs, Zhao Bin introduced the infrastructure direction for real-time intelligent interaction: through a global real-time transmission network, distributed cloud-edge-device joint inference, real-time orchestration, and full-duplex interaction capabilities, models and Agents can continue to collaborate in real scenarios. The Agora conversational AI engine also centers on full-duplex interaction, connecting models, Agent runtime frameworks, and users to support a stable, synchronized real-time conversation experience.

At the same time, Leiphone.com learned that Agora continues to improve conversational AI development tools and development kits for developers, combining capabilities such as SDKs, Skills, CLI tools, and Recipes to help developers build Agents with complete logic and execution capabilities more efficiently.

From customer service and education to AI companionship and smart wearables, Zhao Bin believes that conversational AI is entering more real scenarios. Whether AI can truly create value depends not only on whether the model can "talk," but also on whether it can understand needs, communicate naturally, and continuously complete tasks. Real-time interaction technology will play an increasingly important role in this process.

Subsequently, Andrey Norkin, chief architect at NVIDIA and co-chair of the AOMedia Video Codec Working Group, gave a speech on AV2 video codec technology.

Roundtable: Will personal agents become the next-generation personal entry point?

In the first roundtable hosted by Agora co-founder Tony Wang, interactive media product manager and tech blogger Eugene Wei, tech self-media personality and host of the podcast "Luanfan Shu" Pan Luan, and AIoT hardware entrepreneur and edge AI self-media host Hua Jianxin discussed "Personal Agents: From New Tool to New Entry Point?"

From phones, glasses, and earphones to more wearable devices, personal agents are once again prompting the industry to think about the next-generation human-computer interaction entry point. Pan Luan believes that personal agents may bring a paradigm shift: the value of AI products is shifting from competing for user attention to helping users complete tasks more efficiently. Future personal agents should not only respond to explicit instructions, but also proactively take on tasks and help users handle matters that have clear goals but are not worth investing a large amount of time in.

For next-generation personal hardware, Hua Jianxin proposed that the future entry point may not be monopolized by any single type of device; glasses, earphones, and other hardware capable of perceiving the environment and executing tasks all have opportunities. He believes that a personal agent not only needs to understand the user, but also needs to continuously perceive the real world through physical devices and take action, which brings new development space for hardware of different forms.

Eugene Wei steered the discussion toward the value of "context": whether a personal agent can truly understand the user's interests, communication records, and social relationships will affect whether it can become a valuable personal entry point. In assessing whether a new entry point holds, he also proposed paying attention to users' daily interactions with the Agent, and whether users more often seek out the Agent or the Agent more often seeks out users. The roundtable thus further discussed the product value, hardware forms, and future metrics of personal agents.

In the second roundtable, moderated by Cui Bo, Chief Commercial Officer of Agora, Chen Wei, founder of Xieyue Intelligence, Geng Xianglei, founder of Guanke Technology, Cai Guoxiang, co-founder and ecosystem lead of Rokid, and Wu Kan, CEO of Titan Qiming, explored conversational AI from different scenarios including embodied intelligence, smart wearables, enterprise marketing, and AI recruiting: driving the next generation of human-Agent interaction.

Chen Wei believes that the key to embodied intelligence is truly translating human intent into robot action. Robots not only need to understand voice commands, but also need to fuse multimodal information such as vision, voice, and their own state to complete perception, planning, and decision-making about the physical environment, and to continuously adjust task execution based on environmental changes after actions. Compared with completing preset demonstration tasks, it is more important for robots to continuously understand people's real needs in open environments, so that actions always remain consistent with human intent. Xieyue Intelligence is developing an embodied foundation model under the new Duplex Reasoning paradigm, aimed at achieving integrated robot interaction and action.

Cai Guoxiang, starting from the practice of AI glasses, emphasized that the conversational experience is the core competitiveness of AI wearable devices. He said that no matter how powerful the hardware and the AI capabilities behind it are, if the conversation is not natural enough, the user experience will be directly affected. In the future, AI wearable devices will need low-latency, stable real-time interaction, as well as understanding of semantics and long-term context, so that users gradually no longer feel they are talking to AI. He also shared the practice of the Rokid AI assistant using Agora's RTC link service, and expressed optimism that AI glasses will become an important personal entry point in the AI era.

In enterprise application scenarios, Geng Xianglei pointed out that for AI voice agents to achieve reliable business delivery, they must solve issues such as latency, interruption timing, and human-likeness, among which latency may even directly affect task outcomes. For To B marketing, AI also needs to combine with industry knowledge systems and human collaboration to better safeguard business outcomes. Wu Kan, drawing on AI recruiting practice, said that conversational AI can help recruiting businesses reach job seekers at scale and improve service efficiency, rather than simply replacing humans. To form a business loop, it is also necessary to accumulate industry data, organize the business with job seekers as the subject, help the Agent build long-term memory, and assist AI in continuously evaluating and optimizing service quality around actual business events such as interviews and onboarding.

From the sharing of the four guests, it can be seen that although different industries have different requirements for conversational AI, they point in a common direction: AI must not only be able to communicate, but also understand real scenarios, integrate into existing business processes, and reliably complete tasks. With the continued development of real-time interaction, multimodal perception, and Agent execution capabilities, conversational AI is expected to further evolve from an interaction tool into an important force connecting the digital world and the physical world and driving business operations.

From real-time interaction to real-time intelligence, opening the next chapter of conversational AGI

From the five-level conversational capability framework proposed by Zhao Bin to the two roundtables' discussions on personal agents and real-world business implementation, the iRTE2026 main forum presented a clear industry trend: competition in conversational AI is extending from basic model capabilities to context understanding, real-time interaction, cognition of complex relationships, and task execution.

Looking toward the next stage of conversational AGI, Agora will continue to explore the integration of real-time interaction and generative AI with developers and industry partners, promoting more natural and efficient real-time interaction experiences between people and people, and between people and Agents.