On August 3rd, Alibaba launched QwenWork, integrating QoderWork, MuleRun, and Wukong into a single platform designed to boost productivity using AI. This development is aimed at large organizations rather than individual users. The integration of Alibaba Cloud, Qwen models, and DingTalk allows agents to operate directly within existing messaging (IM), CRM, and ERP systems, eliminating the need for a new working interface.
Initial success metrics have been quite high. In its first month, QwenWork attracted over 30 million users, with corporate accounts exceeding half of that base. On August 17th, Alibaba also made the MyContext project publicly available—an infrastructure that transforms disparate work data, such as emails, documents, chats, and approvals, into structured context for agents. This repository quickly gained thousands of stars on GitHub.
The launch of QwenWork comes amid a phenomenon dubbed 'Token Maxxing' by Chinese operators—the purchase of massive monthly token quotas per employee without a clear link to business results. Similar debates about return on investment are observed in major global tech companies, which are tightening AI usage permissions following experimental spending. The core thesis of QwenWork is that the bottleneck lies in organizational structure, not in the sheer volume of raw tokens needed to complete tasks.
A prime example of application is the automotive conglomerate Changan Automobile. Under the 'No AI, No Changan' initiative, teams are implementing assistants across all departments: R&D, manufacturing, procurement, sales, and service. One engineer created a tool for uploading photos that extracts fields from vehicle dispatch documentation and cross-references more than 100 mandatory points against regulatory requirements. A procurement manager reported AI assistance in 58 different scenarios, amounting to about 240 interactions over two weeks, saving approximately 45 hours. Subsequently, he shared these skills with colleagues, enabling them to immediately adopt most of the functionality.
Other companies that have adopted the solution include research groups compiling quality control cycle reports, logistics divisions using AI for planning and monitoring, and pharmaceutical company finance departments verifying balances and reports. The long-term vision involves the 'AI completion' of IT infrastructure: ERP and OA surfaces become accessible to both humans and agents, transforming personal prompt tricks into corporate assets. QwenWork bets that governance, shared context, and digital colleagues will surpass another round of empty token budgets.