Chen Yusen.

The last time this name appeared in tech news, it was for a series of security moves after his company Chaitin Tech was acquired by Alibaba. On July 1, 36Kr reported that he will lead Alibaba's newly consolidated enterprise productivity AI product -- building on QoderWork as the foundation, integrating the capabilities of Wukong and MuleRun into a more powerful Agent product for enterprise scenarios.

Internal Horse Racing Ends, Big Tech Enters Product Consolidation

Over the past year, Chinese LLM companies and internet giants have generally adopted an "internal horse racing" strategy in Agents: multiple teams simultaneously building Agent products in different directions, with resources going to whoever pulls ahead. Alibaba's consolidation sends a clear signal: the horse racing phase is winding down; resources are being concentrated on the B2B productivity market.

Alibaba responded that QoderWork, Wukong, and MuleRun's existing products and services will see seamless upgrades, with no impact on user benefits. The three products each have their own focus: QoderWork leans toward code development and engineering productivity, Wukong toward general task handling and personal assistance, MuleRun toward workflow automation and enterprise system integration. After consolidation, the three products' capabilities will be unified under a single product framework.

This isn't just Alibaba. Over the past month, ByteDance, Baidu, and Tencent have all been consolidating their internal AI product lines to varying degrees -- the consumer traffic war has largely played out, and everyone realizes: what generates real large-scale revenue isn't $20/month consumer chatbot subscriptions, but productivity transformation in the enterprise market.

Feishu (Lark) also released new AI features last week: an AI assistant that lives directly in spreadsheets, summonable by @mention in group chats, handling all the dirty work. DingTalk and WeCom are doing similar things. Alibaba's consolidation is essentially stockpiling ammunition for the coming enterprise Agent wars.

Three Key Variables in the Enterprise Agent Market

Why are big tech companies going all-in on B2B productivity Agents now? Three reasons.

First, the consumer monetization ceiling is visible. ChatGPT took two years to reach roughly 100 million paying users, but consumer willingness to pay and ARPU have caps, and competition is fierce -- users don't need simultaneous subscriptions to ChatGPT, Claude, Gemini, Kimi, and Doubao. The enterprise market is completely different: a mid-size company might have an AI budget of millions or tens of millions RMB per year, and once embedded in workflows, switching costs are high and renewal rates far exceed consumer levels.

Second, Agent technology has genuinely matured to the point of doing real work. A year ago Agents still regularly crashed in demos; now OpenSquilla and its peers can self-verify and iterate in closed loops, and Feishu's spreadsheet Agent can already handle actual work. Technical maturity has crossed the line between "toy" and "tool."

Third, capital and state capital are entering the arena. On the same day Alibaba announced consolidation, the CETC AI Industry Equity Fund was registered with 2 billion RMB in capital, funded by CETC, Suzhou Xiangcheng Financial Holdings, and the Central Enterprise Strategic Emerging Industries Fund. State-backed industrial funds entering the scene means AI B2B isn't just commercial activity -- it's being driven by policy as well.

Of course, competition in B2B productivity Agents won't be easy. Alibaba faces rivals including: Microsoft Copilot already deeply integrated with Office 365, Feishu and DingTalk with natural entry advantages in enterprise collaboration, vertical SaaS companies (like Salesforce, Notion) embedding AI Agent capabilities, and numerous startups attacking niche scenarios.

But Alibaba's advantages are also clear: Qwen's model foundation, Alibaba Cloud's infrastructure, DingTalk's enterprise customer base, and the accumulated capabilities of three validated products. Putting Chen Yusen in charge also sends a signal -- this isn't a marginalized internal innovation project; it's Alibaba's main force on the AI B2B battlefield.

The enterprise AI market in the second half of 2026 will be even more exciting than consumer. The chatbot wars may be approaching endgame, but the enterprise productivity Agent wars have only just begun.