AI Rapidly Transforms Auto Industry; Companies Must Build a "Knowledge Brain" and "Execution Playbook"

At the High-Level Forum on Intelligent Electric Vehicle Development (AI + Automotive Themed Forum) held on April 12, 2026, Yang Liwei, Vice President of Volcano Engine, stated that the integration of AI and automobiles has moved beyond its initial phase. As of the end of March 2026, the daily token usage of Doubao's large model surpassed 120 trillion, representing a 1,000-fold increase since its launch and capturing the majority share of the Chinese market. Yang noted that AI interaction is shifting from a human-centric chatbot model to a new paradigm centered on AI as the primary workflow driver—where AI handles task planning and execution, while humans assume roles of control and oversight. This trend is already evident in scenarios such as full-duplex voice in intelligent cockpits, AI-powered customer service, automated ad placement, and code generation. He emphasized that a complete AI Agent architecture must include goal setting, knowledge base construction, tool invocation capabilities, and a closed-loop execution mechanism. Regarding enterprise AI adoption, Yang proposed prioritizing the development of two core assets: first, an "AI Knowledge Brain" to codify business rules, system interfaces, and process logic; and second, an "AI Execution Manual," which encapsulates various API capabilities through Skills or MCP (Model-Controller-Protocol) approaches. Additionally, he stressed the need to optimize human-AI collaboration efficiency and related management systems. Furthermore, Yang revealed that Volcano Engine will unveil a new AI-centric human-vehicle interaction solution during the Beijing Auto Show, driving automotive interactions toward greater proactivity and automation.