Alibaba’s Two Secret Teams Merge with Wukong: The Target Is Not WorkBuddy

Huxiao has exclusively learned that Alibaba internally regards "AI Office," currently a market hotspot, as one of its strategic directions in the AI field. Regarding "AI Office," Alibaba's decision-makers have positioned it as an enterprise-grade AI productivity platform oriented toward corporate organizations, rather than merely a personal AI office tool for consumer-facing (C-end) users.

In terms of organizational structure, the core carrier for "AI Office" is the "Qwen Office Business Unit," a Level 1 business unit under the ATH Business Group. This organization was upgraded from the former Wukong Business Unit, which was led by Chen Yusen as the person in charge.

In terms of products, Qwen Office (QwenWork) will be Alibaba's core product in the AI office scenario. This product is not a simple merger or upgrade of existing products but a rebuild based on the technological accumulation of three products: QoderWork, Wukong, and MuleRun (Mule Kuai Pao), achieving a "technical trifecta."

Huxiao has learned that QoderWork, Wukong, and MuleRun (Mule Kuai Pao) will no longer exist as standalone products in the future, and their talent pools have been integrated into the unified Qwen Office system.

Based on these adjustments and changes, the specific structural changes within ATH are as follows:

· Token Foundry Business Unit: Led by Wu Yongming, including the Tongyi large model team and the Future Life Lab (projects such as Happy Horse and Happy Oyster)

· Qwen Business Unit: Led by Wu Jia, Vice President of the Group, responsible for C-end applications such as the Qwen APP

· Qwen Office Business Unit: Led by Chen Yusen, responsible for the two core businesses: Qwen Office and DingTalk

· MaaS Business Line: A model service platform and technical system linking upstream and downstream to connect Token pathways

· Innovation Business Unit: Composed of multiple super-unit groups, conducting innovative applications and incubating new projects

Insiders told Huxiao that senior group executives, including Wu Yongming, personally attend key meetings for Qwen Office. Qwen Office has already become one of the most critical strategic-level products within the company. Enterprise Agents: Who can successfully execute real workflows?

Starting in March 2026, the explosive growth of the "AI Office" market has turned this battlefield into a new "colosseum" for tech giants. Products such as Alibaba's QoderWork and Wukong, Tencent's WorkBuddy, and ByteDance's TRAE IDE underwent high-frequency "iterations and launches" in the first half of 2026.

In terms of traffic, statistics from Analysys show that the total monthly desktop visits for "Office Agents" across the entire industry were approximately 20 million in early 2026. By June, this figure surged, breaking through 60 million.

Data from OpenRouter shows that global Token calls spiked in Q2 2026, with programming and office scenarios contributing more than half of the volume.

However, in this market, Tencent currently holds the first-mover advantage in seizing "market share."

Data from multiple third-party statistical agencies shows that in June, Tencent's WorkBuddy achieved 20.97 million monthly desktop visits for AI office agents; ByteDance's TRAE IDE (China version) ranked second with 12.79 million visits; Alibaba's QoderWork ranked third with 7.88 million visits (with Wukong's 1.31 million visits not included in this count).

In terms of market share, the current ranking of monthly visit shares for Tencent, ByteDance, and Alibaba is approximately as follows:

· Tencent Series:

WorkBuddy + CodeBuddy + QClaw + Marvis totaled 32.62 million visits, accounting for 53.8% of the total market

· ByteDance Series: TRAE IDE + Coze totaled 14.41 million visits, accounting for 23.8%, ranking second

· Alibaba Series: QoderWork + Wukong totaled 9.19 million visits, accounting for 15.2%, ranking third

Huxiao has learned that Alibaba's decision-makers believe the opportunities in the AI office market have just begun and the final outcome has not yet been determined. This means Alibaba still has the opportunity to expand its advantages.

On August 3, Qwen Office officially launched its public beta. The web version and standalone PC client are now available, and built-in entry points for DingTalk on PC and mobile will open soon. On the same day, Alibaba also announced its latest model, Qwen3.8—reportedly further enhancing capabilities in Coding and Cowork.

Two Confidential Projects Form the Foundation

Huxiao has exclusively learned that Alibaba's internal Q-Team and S-Team are the talent and technological foundations of Qwen Office.

Q-Team was established in the first half of 2025, led by Ding Yu (alias: Shutong), Vice President of ATH. The project originated as a "POC exploration targeting developer geeks," making it the earliest team within Alibaba to venture into Coding Agents.

In August 2025, Q-Team released its first product, "Qoder"—which was Alibaba's core product with the largest market share in AI office solutions before the integration into Qwen Office.

Also in the first half of 2025, the S-Team, led by Chen Yusen, emerged. This project was positioned around cloud-based AI Agent capabilities and internally tested its first product, MuleRun (Mule Kuai Pao), in August 2025.

Notably, Q-Team and S-Team were not in a relationship of "internal horse racing." Instead, they functioned more like AI solutions targeting different scenarios and user groups. This led to distinct differences in their subsequent technical strengths:

· Qoder, extended from Q-Team, initially emphasized the developer community, so its subsequent development focus was primarily on AI Coding and model reasoning capabilities;

· MuleRun (Mule Kuai Pao), extended from S-Team, heavily emphasized cloud Agent capabilities and long-chain task execution abilities.

In fact, the first half of 2025 was also when DingTalk began promoting its "AI-driven restructuring." This internal DingTalk project was the predecessor to what later became Wukong.

From a spatiotemporal perspective, the first half of 2025 effectively became the common starting point for these three forces that would later merge.

Huxiao has learned that in the first half of 2025, Alibaba internally observed the agent-oriented trends of Silicon Valley products like Claude, highlighting programming and office capabilities. Simultaneously, based on observations of internal employees' office and programming needs, Alibaba judged that agent-based programming and office capabilities would definitely be the future direction.

Based on this, "super group"-style projects like Q-Team and S-Team emerged within Alibaba. These projects received significant attention from Alibaba's decision-makers, with project leads often needing to communicate directly with senior executives like Wu Yongming regarding progress.

However, Alibaba's problem-solving approach differed from the strategies of ByteDance and Tencent.

From the first half of 2025 to June 2026, over nearly a year, these three products essentially "fermented" independently within their respective scenarios and user groups.

Huxiao has learned that internal integration was considered in 2025, but at that time, the focus was more on cross-product synergy and technology reuse.

The Target Is Not WorkBuddy

The turning point occurred in the first half of 2026.

Alibaba's decision-makers discovered that the three products were showing signs of "convergence"—as "Agent + Office" became the consensus mainstream direction, specific functions and product capabilities began to "converge and evolve" more frequently.

The market was also changing. Starting in Q1 2026, the Agent trend reached its peak, and both consumer and enterprise users began to form some standard consensuses regarding the demands for AI office tools:

· Simpler and easier to use, even "foolproof"

· Output-result oriented, focusing on "deliverable outcomes" rather than entertainment consumption

· More natural and painless embedding into existing office tools and workflows

These factors became the catalysts for the "trifecta" integration.

Huxiao has learned that Chen Yusen, born in 1992, was chosen to lead Qwen Office, partly due to the "Agent" problem-solving capabilities he demonstrated while leading MuleRun (Mule Kuai Pao).

After determining the integration, positioning became a key issue.

In May and June, the business team and Alibaba's decision-makers held multiple meetings to discuss, finally reaching a consensus: Alibaba's opportunity lies in the "enterprise-grade" sector.

An insider told Huxiao that enterprise-grade AI office solutions must ensure robust general Agent capabilities while also being able to gain insights into and meet enterprise-level scenario requirements.

This individual believes that while many vendors can achieve the former—with Tencent, ByteDance, and Baidu all possessing this capability—Alibaba's advantage in this area lies in its Qwen large model strength and "engineering" capabilities. Regarding the latter, since Alibaba has accumulated millions of enterprise customers through DingTalk and Alibaba Cloud, its vast resources of demand data and scenario resources will become Alibaba's trump card in the AI office sector.

Furthermore, Alibaba's decision-makers have granted the Qwen Office team considerable patience.

Huxiao has learned that in the short term, Qwen Office will not prioritize metrics such as "number of enterprise users" or "monthly active users" as the most important assessment targets. Instead, priority will focus on two other dimensions:

· Product usability: Primarily assessing whether the two core capabilities, "General Agent Capability" and "Enterprise Scenario Functions," are becoming increasingly mature and usable, pursuing continuous optimization of the user experience.

· Insight and implementation of enterprise scenario needs: Primarily assessing the depth of understanding and insight into real enterprise needs, and the ability to design truly valuable product features based on these needs, as well as executing effective enterprise-grade implementation plans.

However, challenges remain.

Insiders revealed that the team is urgently polishing industry customer benchmarks and enterprise customer benchmarks—as an enterprise-grade product, Qwen Office urgently needs successful cases as demonstrations to attract more enterprise users.

Moreover, ByteDance and Tencent are also eyeing the enterprise-grade AI office opportunity.

In June, Tencent released the enterprise version of WorkBuddy, formally entering the enterprise market; in July, ByteDance underwent its largest-scale adjustment in its ToB business history, merging the Feishu product entirely into Doubao, and integrating Feishu's sales, marketing, and customer success (GTM) teams into Volcano Engine to establish a "Creativity Service Platform" for unified ToB delivery.

The author believes that this race surrounding AI office solutions is currently akin to a 3,000-meter final where the starting gun has just fired and participants have only run 100 meters. In the next six-month cycle, it will be a critical period for competing on product and service quality to win market mindshare.

The core challenges facing Alibaba include, but are not limited to:

· How to more rapidly complete the full integration of technology and products for the "trifecta," and significantly improve the Agent feel of Qwen Office

· How to effectively leverage DingTalk's existing advantage in enterprise customers for Qwen Office, which may involve not just Token consumption and scenario development, but also more ingenious FDE (Field Development Engineer) systems

· There is a significant difference between the mindset of enterprise customers using AI office tools and that of individual consumers; how can Qwen Office enhance the sense of gain for enterprise customers?

The good news is that the market has reacted positively to Alibaba's launch of Qwen Office.

As of the midday trading session in Hong Kong stocks on August 3, Alibaba's shares rose by more than 7%, with its total market capitalization rebounding to approximately HK$2.4 trillion.

Risk Warning and Disclaimer

The market involves risks; investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment decisions made based on this content are the sole responsibility of the investor.