Is the “entrance battle” for AI offices over?

shadow memo

2026.11.20 09:46

· From Guangdong

Full text 5018 words

Text | Shadow Memo

On November 18th and 19th, Tencent and Byte handed over their respective report cards on the AI ​​office track almost simultaneously.

Tencent WorkBuddy's daily active users have exceeded 10 million, the average daily Token calls have exceeded 100 trillion, and the annualized revenue in the past seven days has exceeded 500 million US dollars, which is more than twice that of Tencent Conference.

At almost the same time, Byte's Doubao work also announced the latest progress: in just three months after its launch, DAU exceeded 8 million, and enterprise customer ARR exceeded 2 billion yuan.

On the other hand, Alibaba Qianwen Office is also accelerating to catch up. Analysys analysis data shows that the user penetration rate of WorkBuddy reaches 55%, and the high-frequency usage ratio reaches 92%; Qianwen Office follows closely, with the high-frequency usage ratio reaching 88%.

If you only look at these numbers, it is easy to draw a conclusion: China's AI office market is ushering in a new round of traffic war.

But in terms of observation, what really deserves attention in these data is not who has higher daily activity, nor whose income is growing faster, but that the three Internet giants Tencent, Byte, and Alibaba are using three completely different business logics to re-compete for a hugely valuable market.

Tencent hopes to make AI a work entry point for everyone, Byte hopes to make AI a productivity infrastructure for enterprise organizations, and Alibaba is trying to use DingTalk’s corporate ecosystem to deeply embed AI into organizational collaboration systems.

This means that the competition in AI offices has entered a new stage.

In the past, everyone competed to see whose model was smarter. Now everyone competes to see who can truly take over the user's work.

Tencent and Byte showed their transcripts at the same time

Let’s look at Tencent first.

The data released by WorkBuddy this time actually has three dimensions that are most worthy of attention.

The first is the scale of users, with daily active users exceeding 10 million, indicating that AI office is no longer just an exclusive tool for technology enthusiasts and programmers, but has begun to enter a wider range of daily work scenarios.

The second is the depth of usage, with the average daily Token calls exceeding 100 trillion. Token consumption reflects the frequency of user interaction with AI and the complexity of tasks to a certain extent. Compared with simple downloads and registrations, it can better reflect whether the product is truly integrated into the work process.

The third is commercialization capabilities. The annual revenue in the past seven days has exceeded US$500 million, and it is already more than twice that of Tencent Conference.

This set of data releases an important signal: AI office is moving from the traffic acquisition stage to the commercial value verification stage.

In the past, the core indicators of Internet products were user size and usage time. But for AI products, just having users is not enough.

Whether users are willing to continue calling models, whether they are willing to hand over complex tasks to AI, and whether they are willing to pay for a more efficient work experience determines whether a product can form a true commercial closed loop.

Tencent WorkBuddy obviously hopes to use these three indicators to prove that it not only has traffic, but also usage depth and monetization capabilities.

Look at the bytes again.

Three months after Doubao was launched, DAU exceeded 8 million, and corporate customer ARR exceeded 2 billion yuan.

Unlike Tencent, which emphasizes individual user scale, token calls and income elasticity, the signal released by Byte this time is more biased towards the enterprise market.

ARR, or annual recurring revenue, is an important indicator of the commercialization capabilities of subscription-based enterprise software. The willingness of enterprise customers to sign up means that AI has begun to transform from an early adopter tool into a productivity product that enterprises are willing to continue to purchase.

What is particularly noteworthy is that such results have been achieved within three months of going online, indicating that the demand for AI office work among enterprises is rapidly increasing.

Of course, Tencent's seven-day annualized revenue and Byte's ARR are not exactly the same statistical caliber. You cannot simply put the two sets of numbers together to determine who has stronger commercialization capabilities.

But together they prove one thing:

AI office is no longer just a story about the future told by Internet giants, but has begun to generate real business income.

And this is precisely the most important change in the entire industry.

Tencent: Occupy the entrance first, then look for room for monetization

Why was Tencent WorkBuddy able to gain such a huge user base in a short period of time?

The answer lies largely in Tencent’s ecosystem.

For Tencent, the biggest advantage of AI office is not necessarily the model itself, but the huge user base represented by WeChat and the natural traffic entrance formed by it.

Over the past decade or so, WeChat has become an important infrastructure for Chinese Internet users to connect work and life.

From personal communication and file transfer, to group chat collaboration, public account content acquisition, to mini programs and corporate WeChat, Tencent has actually mastered a large number of information touch points in users' daily work.

The emergence of AI office has given these touch points new imagination space.

Just imagine, a user used to receive files in WeChat, then open office software to process documents, and finally send the results to colleagues through chat tools.

In the future, this series of actions may be reorganized by AI.

Users only need to put forward requirements, and AI can help them understand documents, refine information, generate content, and even assist in completing subsequent tasks.

This is what Tencent hopes to achieve: turning AI from an independent tool into an assistant that users can call at any time during their work.

From a business model perspective, this is a typical entry-level AI route.

Its core is not to first persuade companies to purchase a complete set of software, but to first allow enough individual users to form usage habits, and then expand business value through high-frequency use, value-added services and enterprise scenarios.

This logic is a bit like WeChat in the past.

First establish user connections, and then gradually extend business scenarios such as payment, advertising, mini programs, and enterprise services.

Of course, AI office does not simply copy the success of WeChat.

Because the core value of chat tools is to connect people, and the core value of AI office is to help people complete tasks.

This means that users will not use WorkBuddy for a long time just because it has the Tencent ecosystem behind it.

What really determines user retention is its ability to understand work content, complete complex tasks, and continue to reduce user work costs.

Therefore, Tencent's most important task at present is to transform its traffic advantage into work ability advantage.

If you just have a large number of users but cannot penetrate into the users' real workflow, then the entrance may be just a traffic entrance instead of a productivity entrance.

Byte: Integrate AI into workflow and let companies pay for efficiency

Unlike Tencent, Byte's AI office route is more focused on organization and process.

The important advantages behind Doubao's work are Byte's accumulation in scenarios such as enterprise collaboration, document and knowledge management, and the enterprise customer base formed by Feishu.

The biggest difference between corporate office and personal office is that individuals can change tools because they are easy to use, but it is difficult for companies to completely change the original organizational system just because a new tool appears.

An enterprise may have hundreds or even tens of thousands of employees, and their daily work involves documents, meetings, approvals, project management, knowledge bases, and internal communications.

There are a large number of data relationships and process connections between these tools.

If AI is just an independent chat window and employees need to constantly copy files, upload information, and switch applications, it will be difficult for it to truly change the work efficiency of the enterprise.

But if AI enters documents, meetings, knowledge bases, and collaboration processes directly, the situation will be completely different.

It can generate minutes based on meeting content, answer business questions based on the corporate knowledge base, organize tasks based on project progress, and even help employees complete some repetitive process work.

AI is no longer just a tool waiting for users to ask questions, but has become a participant in the enterprise workflow.

This is the core value of process-based AI.

Why does Byte pay special attention to enterprise customer ARR? Because there are obvious differences between the business logic of enterprise software and personal Internet products.

Individual users may use AI every day, but may not be willing to pay; but as long as companies can confirm the efficiency improvements brought by AI, they will be motivated to include it in their annual procurement budgets.

For enterprises, what is really important is not how many words employees chat with AI every day, but whether AI can reduce duplication of work, shorten business cycles, and reduce management costs.

Therefore, Doubao Work announced that the ARR of enterprise customers exceeded 2 billion yuan. Its significance lies not only in the scale of revenue, but also in proving that AI is entering the enterprise's software procurement system.

Of course, ARR does not mean that all customers have achieved deep usage, nor can it be directly equated to the revenue that has been realized.

But at least it shows that the enterprise market has begun to pay for AI productivity.

From this perspective, Byte is not betting on a smarter chatbot, but an AI work system that can be embedded in enterprise organizations.

Why can't Ali be absent?

In addition to Tencent and Byte, Alibaba is also a player that cannot be ignored in the AI ​​office battlefield.

Qianwen Office’s advantages come from Alibaba’s long-term accumulated enterprise service capabilities and the organizational collaboration network established by DingTalk.

If Tencent is better at connecting individual users and Byte emphasizes the combination of AI and workflow, then Alibaba’s advantage lies in the corporate organization itself.

DingTalk has penetrated into the daily management scenarios of a large number of enterprises, from attendance, approval, meetings, to project collaboration and corporate communication, and has accumulated a large number of organizational-level application scenarios.

These scenarios are naturally suitable for AI transformation.

For example, in the past, employees had to manually fill in approval information, search for corporate systems, and organize meeting minutes. In the future, AI can directly understand the information and processes within the company and help employees complete corresponding tasks.

More importantly, Alibaba has enterprise-level cloud computing and AI infrastructure capabilities.

When enterprises begin to deploy AI on a large scale, model capabilities, computing resources, data security and system integration will all become important factors in purchasing decisions.

This makes Alibaba's competition not just a competition for office software products, but a competition that extends from the office entrance to enterprise digital infrastructure.

Analysys analysis data shows that the high-frequency usage ratio of Qianwen Office has reached 88%, which is not far behind WorkBuddy’s 92%.

This at least shows that the AI ​​office market does not have a dominant position.

The three giants are all looking for their own niche.

What Tencent wants to solve is how to turn individual users into long-term paying users; what Byte wants to solve is how to transform AI capabilities into enterprise-level productivity; what Alibaba wants to solve is how to use organizational and infrastructure advantages to promote the large-scale implementation of AI within enterprises.

Their competitive paths are different, but they ultimately point to the same goal: becoming a truly indispensable work entrance for users and enterprises.

Major manufacturers begin to redefine AI office

There is another detail worth paying attention to.

In the past period, Tencent, Byte and Alibaba have all launched a number of AI office-related products internally, exploring the market through different teams and different product forms.

However, as the market gradually matures, the three major manufacturers have invariably begun to concentrate the capabilities and resources of various internal AI products on core brands such as WorkBuddy, Doubao Work, and Qianwen Office.

Behind this is actually a very important industry change.

AI office is no longer a market that can be won by relying on single-point innovation.

In the early days, everyone paid more attention to model capabilities, generation effects and product experience. Whoever can launch a useful AI assistant first will likely get the first batch of users.

But now, market competition is starting to get complicated.

Because a real AI office product needs to have model capabilities, product experience, user portal, data connection, workflow integration and commercialization capabilities at the same time.

These capabilities are difficult to accomplish by multiple separate products.

Especially in the enterprise market, customers do not want to purchase five or six AI tools with similar functions, nor do they want employees to repeatedly switch between different applications.

What they need is a unified AI working system.

Therefore, large manufacturers concentrate resources on core brands, which is essentially shifting from the technology exploration stage to the market operation stage.

In the past, the competition was about who could make a better AI product. Now the competition is about who can build a complete user, technology and business ecosystem around a core product.

This also means that the competitive threshold for AI office is rapidly increasing.

The ultimate outcome of AI office is to redefine the work entrance

So, among Tencent, Byte and Alibaba, who has the better chance to become the winner of the next generation of AI office?

As far as the current situation is concerned, it is too early to draw conclusions.

Because the data currently released by the three companies represent different business dimensions, user scale, token call volume, high-frequency usage and enterprise ARR cannot be directly equated.

More importantly, AI office still has a problem that has not yet been truly solved:

When can personal workflow and corporate workflow be truly connected?

This is the key to determining the industry structure in the next stage.

Give an example.

An employee handles both individual tasks and team projects every day. He needs to read emails, organize documents, attend meetings, query the corporate knowledge base, and collaborate with other departments.

If AI can only process personal files, it cannot truly understand corporate business; if AI can only run in corporate workflows, it will be difficult for it to cover the large number of fragmented personal work needs of employees.

A truly valuable AI office system must understand the working relationship at the three levels of individual, team and enterprise at the same time.

It not only knows what the user is doing, but also knows what the user's team is promoting, and it also knows what data, systems and business processes the enterprise has that can be called upon.

This is the real competitive barrier for AI office.

From this perspective, Tencent's advantage lies in personal user portals, Byte's advantage lies in workflow and enterprise collaboration, and Alibaba's advantage lies in enterprise organization and infrastructure.

But all three companies need to complete the same puzzle: how to form a real connection between personal AI assistants and enterprise AI systems.

Whoever completes this step first will have the opportunity to upgrade from an AI office tool to a next-generation work operating system.

And this will also change the business logic of traditional office software.

过去,用户打开办公软件,是为了使用某个具体功能。写文档打开文档软件,开会打开会议软件,管理项目打开项目管理工具。

未来,用户可能不再关心具体应该打开哪个软件。

他只需要告诉AI自己要完成什么任务,AI便可以根据权限和业务上下文,调用不同工具、读取不同数据,并组织整个执行过程。

办公软件依然存在,但它们可能逐渐退居后台,成为AI调用的能力组件。

真正站在前台的,是理解用户意图、掌握工作上下文,并能够调动各种工具完成任务的AI入口。

所以,AI办公的终局不是替代办公软件,而是重新定义办公软件与用户之间的关系。

过去,人找软件完成工作;未来,AI理解人的需求,组织软件完成工作。

这才是腾讯、字节和阿里真正争夺的东西。

从WorkBuddy的千万级日活,到豆包工作的企业ARR突破20亿元,再到千问办公加速追赶,中国AI办公市场已经走过了最初的产品尝鲜阶段。

接下来的竞争,不会只发生在模型排行榜上,也不会只发生在用户增长曲线上。

它将发生在每个人每天打开电脑的第一分钟,发生在企业每一次会议、每一份文档和每一个业务流程之中。

谁能够成为用户真正的工作入口,谁就有机会掌握下一代办公生态的主动权。

这场战争,其实才刚刚开始。

敬原创,有钛度,得赞赏

根据《网络安全法》实名制要求,请绑定手机号后发表评论

快报

09:55

国内推出首个智能座舱AI分级评价体系

09:54

财政部在香港成功发行美元主权债券

09:44

网易港股跌超18%

09:43

港股黄金股集体调整,赤峰黄金跌近7%

09:41

先进封装概念逆势活跃,华微电子涨停

09:37

习近平会见印度尼西亚总统普拉博沃

PCB概念局部活跃,中京电子逼近涨停

09:36

港股大模型股集体下挫,MINIMAX跌超10%

09:34

华为算力概念局部回暖,真视通4连板

算力芯片概念延续调整,天普股份2连跌停

09:30

外国企业对特定运动鞋提起337调查申请

09:29

A股开盘:三大指数集体低开,光通信、黄金、GPU板块明显调整

09:25

央行今日开展3795亿元隔夜逆回购操作

09:22

港股开盘:恒生指数跌1.57%,恒生科技指数跌2.95%

09:10

国家邮政局:前10月邮政寄递业务量、收入双增长

09:06

现货白银日内跌幅扩大至2%,报41.99美元/盎司

MSCI新兴市场股票指数较6月高峰下跌20%

09:05

11月LPR报价出炉:1年期和5年期以上利率均保持不变

09:01

国内商品期货多数收涨,焦炭、焦煤涨超3%

富时中国A50指数期货开跌1.31%

扫描下载App