Alibaba Merges Three Products into Qwen Office, Targeting Enterprise AI Market
Alina Collins
Alibaba merged three AI workplace products into QwenWork (千问办公), launching public beta on August 3 — not as a consumer tool but as an enterprise-grade AI productivity platform. This is Alibaba's core bet to close a gap from third place in China's AI-office race.
What exactly got merged?
Three products are now one: QoderWork (AI coding), Wukong (DingTalk's AI overhaul project), and MuleRun (cloud-based AI agents). None will continue as a standalone product.
This is not an upgrade of any single tool — it is a ground-up rebuild using all three products' accumulated technology.
In plain terms = three separate screwdrivers have been welded into one multi-function tool.
Who built the technical foundation?
QwenWork's tech roots trace to two classified internal teams: Q-Team (AI coding and model reasoning, led by ATH VP Ding Yu) and S-Team (cloud agents and long-chain tasks, led by Chen Yusen).
The two teams were not an internal horse race — they tackled different scenarios. This means → the merger was a proactive consolidation after the products converged, not a Darwinian cull.
The merged unit is the QwenWork Business Division under Alibaba's ATH group, headed by Chen Yusen, born in 1992 — a signal that Alibaba's leadership ranks AI-office work as a company-level strategic priority.
Where does Alibaba rank right now?
June 2026 desktop monthly visits for AI-office agents: Tencent (WorkBuddy, CodeBuddy, etc.) 32.62 million, 53.8% share; ByteDance (TRAE IDE, Coze) 14.41 million, 23.8%; Alibaba (QoderWork, Wukong) 9.19 million, 15.2% — third place.
This means → Alibaba's traffic is less than a third of Tencent's. The gap is real.
But Alibaba's internal view: the market has barely started; the endgame is nowhere close. In the short term, enterprise user counts and MAU are not the top KPIs — priority goes to maturing general-agent capability and landing real enterprise use cases.
Why does Alibaba think it can catch up?
Two trump cards: first, the Qwen large model is competitive on engineering capability (the Qwen 3.8 model released the same day further sharpens coding and collaboration); second, DingTalk + Alibaba Cloud together serve over a million enterprise clients — a deep pool of demand data and scenario resources seen internally as a potential edge.
In plain terms = Alibaba's logic is: we may trail on consumer traffic, but we already have a massive enterprise customer base — pipe the AI capability straight into it.
Teams are working intensively on industry and enterprise reference cases, using showcase projects to draw more enterprise users in.
What are the rivals doing?
June: Tencent launched WorkBuddy Enterprise Edition, formally entering the enterprise tier.
July: ByteDance carried out a major B2B restructuring — Feishu folded entirely into Doubao, and the GTM team merged into Volcengine.
This reflects all three giants placing their chips on enterprise AI-office work almost simultaneously. Whether QwenWork can leverage DingTalk's enterprise base to carve out a differentiated path is the pivotal test in this race.
Content is for reference only, not financial advice.