The AI-driven office solutions market is booming, prompting ByteDance to undertake significant internal restructuring.

On July 30, 2026, Kechuangban Daily reported that ByteDance has revamped its AI business organizational framework to boost collaboration among Doubao, Feishu, and Volcano Engine in enterprise productivity environments.

The Feishu and Doubao product teams are set to merge into a unified Doubao product team, spearheaded by Doubao’s leader Zhao Qi, with Feishu’s head Xie Xin reporting to Zhao Qi. This signals Doubao’s central role post-merger, with Feishu’s future trajectory focusing heavily on AI integration.

Furthermore, Feishu’s GTM (Go-To-Market, including marketing, sales, and customer service) team will merge with the Volcano Engine team to create a new ToB GTM unit dubbed the “Creativity Service Platform.” This entity, led by Volcano Engine’s head Tan Dai, will handle marketing, sales, and customer service for ByteDance’s cloud services.

Doubao, Feishu, and Volcano Engine have traditionally operated as distinct business units. ByteDance’s decision to integrate these three reflects the burgeoning AI office sector’s potential. By fortifying links between underlying models, office entry points, and cloud services, ByteDance aims to bolster the competitiveness of its AI office products, positioning them as formidable rivals to Tencent and Alibaba’s offerings.

Large Model Commercialization Hits Snags, AI Office Emerges as Key Growth Area

Since ChatGPT ignited the AI craze, large models have become a focal point for tech firms. However, the high costs associated with computing power investment, model training, and daily operations have hindered AI assistants from achieving a self-sustaining commercial loop.

For instance, LatePost calculations based on Volcano Engine’s public API prices revealed that in May 2026, Doubao was incurring daily costs in the tens of millions of yuan, while its first-half revenue remained below one million yuan.

This discrepancy stems from AI assistants’ initial positioning as chatbots, offering only Q&A services. Unlike search engines, which generate advertising revenue through bidding rankings, AI assistants have struggled to monetize by creating differentiated user experiences.

Yet, AI assistants’ capabilities extend far beyond this. With the advent of Chain of Thought (CoT) and enhanced abilities like perception, memory, planning, and tool usage, large models have evolved into Agents capable of executing specific tasks.

Against this backdrop, AI firms like Anthropic and OpenAI have launched AI Coding products such as Claude Code and Codex, amassing a vast user base and substantial revenue.

Semi Analysis reports that by May 2026, Anthropic’s annualized recurring revenue (ARR) had surged past $44 billion, quadrupling from approximately $9 billion at the end of 2025.

Similarly, as of July 22, 2026, Codex and ChatGPT Work boasted a combined user base of 10 million. The Information notes that OpenAI’s revenue reached approximately $5.7 billion in Q1 2026, roughly $1 billion higher than Anthropic’s. OpenAI’s Chief Revenue Officer, Denise Dresser, disclosed that the company’s ToB business accounted for over 40% of its revenue.

Anthropic and OpenAI’s success in commercializing AI Coding products stems from their focus on productivity scenarios. Unlike the fragmented needs of individual users, enterprises in productivity settings prioritize “cost reduction and efficiency enhancement.” AI Coding products offer benefits like compressed development cycles, automated vulnerability screening, and rapid prototype verification, creating tangible value and naturally persuading ToB users to pay.

However, while AI Coding’s commercialization path is relatively mature, not all enterprises require programming, limiting its potential customer base.

Consequently, many tech firms are extending AI Coding’s commercialization path to explore more universal ToB office scenarios, aiming to further unlock Agents’ commercial value by meeting enterprises’ daily office needs.

Three Business Lines’ Synergy Fuels ByteDance’s ToB Business Integration

On the surface, AI office products’ commercial value lies in boosting ToB users’ willingness to pay by enhancing productivity. In reality, due to the deep integration of cloud, models, and data technologies, AI office products also hold significant strategic value.

As Alibaba Cloud’s Vice President Liu Weiguang observed, “In the cloud computing era, our business model was relatively straightforward, but a longstanding pain point persisted: when reviewing clients’ IT budgets, we couldn’t tap into the portions allocated to internal software development and outsourced labor. Now, it’s the opposite—these budgets are precisely what AI Coding can fully capture.”

Previously, cloud service providers could only offer computing resources like servers, storage, and databases, requiring downstream clients to hire their own programmers or outsourced teams for development. Now, AI Coding can significantly enhance software development efficiency, enabling cloud service providers to enter the massive software development market.

In essence, AI office products not only generate direct revenue but also deeply integrate with cloud services, becoming a vital part of cloud service platforms’ production infrastructure and unlocking new growth opportunities for cloud service providers.

This explains why, beyond Doubao’s efforts to foster deep synergy among Doubao, Feishu, and Volcano Engine in enterprise productivity scenarios, cloud service giants like Tencent and Alibaba are also making similar moves.

On July 20, 2026, Lei Feng Network exclusively reported that Tencent had restructured its organizational framework, transferring the QClaw product center’s related businesses and some teams to Cloud Product Division VI. Cloud Product Division VI manages another AI office agent, WorkBuddy.

On July 27, Alibaba unveiled a one-stop AI office platform called “Qianwen Office,” integrating three products: QoderWork, Wukong, and MuleRun. The platform is led by Chen Yusen, the newly appointed CEO of DingTalk, who previously served as Vice President of Alibaba Cloud and developed the AI Agent product MuleRun.

Overall, major internet firms now recognize the vast potential of AI office solutions and are committed to building full-stack solutions encompassing entry points, data, and cloud services.

However, it’s crucial to note that due to variations in ecological foundations, technical capabilities, and commercial scenarios, the specific development paths and competitive advantages of AI office businesses differ among companies.

AI Office Battle Commences: Tencent, Alibaba, and ByteDance Flex Their Muscles

Analysys data reveals that in June 2026, total visits to 17 mainstream desktop-based AI-native office agent platforms in China exceeded 60 million.

Among them, Tencent-affiliated platforms (WorkBuddy, CodeBuddy, QClaw, Marvis), ByteDance-affiliated platforms (TRAE IDE domestic version, Kouzi), and Alibaba-affiliated platforms (QoderWork, Wukong) ranked in the top three, with visits of 32.62 million, 14.41 million, and 9.19 million, respectively.

Tencent’s core strength in the AI office sector lies in its possession of national-level social traffic entry points, enabling it to swiftly reach a massive user base and enterprises. Social software like WeChat and Enterprise WeChat can not only drive traffic to products like WorkBuddy but also facilitate remote control through account binding, building a deep ecological moat.

However, Tencent’s weaknesses are also apparent. Due to the lack of a powerful self-developed model, WorkBuddy relies on external models such as DeepSeek, Hunyuan, and GLM.

As Li Nan remarked, “Now you can use Kimi K3 in WorkBuddy, but its performance isn’t on par with Kimi Work.” This is because the native capabilities of the base model require supporting Harness, Sub-agent scheduling protocols, and context engines to be fully unleashed. Long-term reliance on third-party models will also cap WorkBuddy’s capability potential.

In stark contrast to Tencent’s relatively weak self-developed model capabilities, Alibaba boasts a high-performance Qwen large model and actively open-sources it. This not only attracts a massive number of developers and lowers the threshold for building Agents but also naturally suits strongly regulated clients in finance, government affairs, and central enterprises.

Additionally, Alibaba’s extensive experience in enterprise data security and organizational collaboration, gained through its longstanding presence in the enterprise service industry, enables its AI office solutions to alleviate enterprises’ data security concerns by ensuring critical data remains within the intranet.

ByteDance’s strength primarily lies in the deep coupling of its office products with large models. Unlike traditional office software that adds AI capabilities to an existing architecture, modules like Feishu Documents, Meetings, and Multidimensional Tables are inherently information flow carriers, making them naturally suitable for Agents to read context.

On this basis, ByteDance’s integration of Doubao and Feishu allows model capabilities to be seamlessly embedded into office suites, making it more appealing to knowledge-intensive SMEs, internet, and tech firms compared to attaching generic third-party models.

Overall, while Tencent-affiliated AI office products currently lead in user visits, the competitive landscape among major players remains fluid. This is primarily due to the rapidly surging demand for AI office solutions, with Alibaba and ByteDance’s organizational restructuring expected to enhance the competitiveness of products like Qianwen Office and Doubao.

For Tencent, Alibaba, and ByteDance, the AI office competition transcends mere product battles; it’s a long-term contest centered on next-generation human-computer interaction interfaces, enterprise data assets, and cloud service ecosystems.

Looking ahead, while traffic entry points and model performance may offer temporary advantages, the real key to victory lies in AI firms’ ability to deeply understand the ToB market’s real pain points and transform Agents into a productivity foundation that can be integrated into business processes, continuously unlocking organizational efficiency.