Alibaba's Qwen team explores personal AI agents with model, memory, and ecosystem synergy

At the 2026 China Computer Conference on October 22, Alibaba's Qwen AI algorithm head Jiang Guanjun outlined the technical approach to developing personal AI agents. He emphasized that true personal intelligence should deliver incremental value and personalized proactive services, relying on three key elements: models, memory, and ecosystem. Jiang categorized AI application development into three stages—dialogue, agent, and assistant—and noted the current shift toward the assistant phase, where AI systems autonomously plan and execute tasks based on user goals and context. He highlighted that the theoretical limit of personal AI agents in solving daily problems is nearly 100%, but unlocking this potential requires aligning system capabilities with problem-solving paths through efficient synergy of models, memory, and ecosystem. Jiang also detailed Qwen's focus on task stability, memory management, and ecosystem integration to enhance user experience across various domains.