By Bishan

Source: Bowang Finance

Four months ago, Xin Xie stood on the podium at the Peking University Entrepreneurship and Innovation Center and delivered a talk titled How Did I Connect the Dots?.

As CEO of Feishu, he had led the business for a full decade. In that speech, he reflected on the "dots" he had personally connected throughout his career: from CTO of Kuxun to ByteDance's first Head of HR, and finally to the creator of an office productivity tool. Those once-scattered dots had now formed a line leading to Feishu. What he didn't elaborate on was where this line would lead next.

At 9:30 a.m. on July 30th, the answer arrived in the form of a company-wide email. The Feishu product team merged with the Doubao product team to form a new Doubao product team, led by Doubao's head Zhao Qi. Xin Xie now reports to Zhao Qi. Meanwhile, Feishu's marketing, sales, and customer service teams were integrated into Volcano Engine, forming a new organization called the "Creativity Service Platform," led by Volcano Engine's head Tan Dai. Feishu's existing products and services remain unchanged.

Brief words, major implications. This once-independent BU, which stood alongside Douyin and Volcano Engine, was now split in two: product capabilities went to Doubao, while commercialization front-end operations went to Volcano Engine. Xin Xie's reporting line shifted from ByteDance CEO Liang Rubo to Zhao Qi. For the first time in a decade, he no longer independently led a business unit.

It should be clarified that only Feishu was dismantled. Volcano Engine served as the recipient in this email, taking over Feishu's entire commercialization front-end.

01

The First Phone Call

The story began with a phone call twelve years ago.

On September 2, 2014, Zhang Yiming called Xin Xie and asked if he would consider switching jobs to join ByteDance as Head of HR. Xin Xie declined on the spot. At the time, he was CTO of Kuxun Travel, with a bachelor's and master's degree in computer science from Peking University. He had served as a project manager at Microsoft Research Asia, a core engineer for Baidu's web search team, and even appeared on Tianjin TV's Only You to recruit for his company—a true technical manager. "I'm the CTO now. Having me join such a small company as HR seems like a big step down," he later recalled thinking.

Zhang Yiming didn't give up. He explained why he insisted on hiring a CTO for the HR role: "Strategy is important, but the team is equally crucial. I've learned this firsthand. Kuxun and Qunar competed in the same direction, but the gap kept widening." Three months later, Xin Xie joined. At the time, ByteDance had around 300 employees, and he became the company's first Head of HR. His first task was to map out the entire organizational structure using tools—something no one had truly clarified before for those 300 reporting relationships.

As HR, Xin Xie quickly turned his attention to tools. In 2014, he said at the QCon conference: "When you use advanced tools, you're not just using the tools themselves—you're unconsciously adopting advanced work methods." For several years, ByteDance's internal office tools changed annually, switching from Skype to WeChat Enterprise, then to Slack, and finally to DingTalk—Feishu later officially stated that ByteDance had once been DingTalk's largest customer. The breakup occurred around 2016: ByteDance proposed customization requests to DingTalk that weren't met, and senior leadership disagreed with DingTalk's manager-oriented product philosophy. They believed good efficiency tools should serve employees.

In 2016, Xin Xie sent Zhang Yiming an email proposing that the company, now with 1,200 employees, develop its own office tools. By year-end, a ten-person team launched Lark, with Xin Xie personally writing the product specification—using GTD principles for instant messaging, treating communication content as documents, and designing workflows that could be checked off like read emails. This product later received a Chinese name: Feishu.

02

The Sharpest Years

Feishu grew organically within ByteDance to address a fundamental question of the digital age: how could information flow more freely within an organization?

It solidified ByteDance's proven organizational practices—transparent information, document-driven workflows, distributed decision-making, and OKR alignment—into a product later summarized as "Context, not Control." In 2019, Feishu opened to external users while launching an international version. By 2021, as ByteDance restructured into BUs, Feishu stood alongside Douyin, TikTok, and Volcano Engine as one of six independent BUs, with Xin Xie reporting directly to Liang Rubo. That marked Feishu's peak within ByteDance.

Those years were also Feishu's sharpest phase. "Advanced teams" like Li Auto, Xiaomi, and NIO became clients, choosing Feishu not just for its product quality but also to gain insights into "how ByteDance works." At early Future Infinity conferences, the main speakers were clients from various industries discussing organization, collaboration, and reducing meetings—seemingly about software, but actually spreading a belief system about how companies should operate.

After reaching its peak, the trajectory mostly pointed downward. By late 2022, company-wide layoffs hit Feishu, cutting over 1,000 employees. At ByteDance's 11th-anniversary event in March 2023, Liang Rubo made a comment later frequently cited: Feishu and Volcano Engine's R&D investment was "no less than Douyin or TikTok," but their "ROI wasn't very cost-effective." The comment had a second half—"Doing enterprise services well is important for the company to become a great tech company"—but the first half dominated discussions. Years later, it's clear that financial considerations outweighed strategic value.

In March 2024, Xin Xie sent a company-wide letter announcing another round of layoffs, reducing the 5,000-person team by about 20% to "return to a startup's Day 1 mentality." Three months later, Feishu President Zhang Nan stepped down, and the president role was abolished. In May 2025, Feishu People business was spun off. By September 2025, the real foreshadowing (foreshadowing) emerged: Doubao's head Zhao Qi joined, and Feishu product head Tong Yao began reporting to him with a dotted line. Ten months later, the dotted line became solid.

03

Why Now, of All Times

The confusion lies here: Feishu's restructuring came precisely during its best performance period.

A July report from Caijing provided numbers: Feishu's 2025 revenue exceeded 3 billion yuan, up ~60% YoY. In Q2 2026, revenue grew over 100% YoY. Under the ARR metric, it reached $100 million in 2022 and doubled to over $200 million in 2023. More critically, over 90% of new clients also purchased Feishu's AI products—enterprise clients had voted with their feet, coming not just for office software but for AI.

Growth was real, and so was pressure. With a ~4,000-person team, industry estimates put annual labor costs between 3.2–4 billion yuan. Xin Xie had once extended profitability goals to 2030. Relying solely on collaboration licensing fees for documents, spreadsheets, and meetings, even with rising ARR, would struggle to cover escalating labor, R&D, and new AI computing costs. Meanwhile, competitors multiplied: DingTalk and WeCom squeezed the existing market, while startups like Moonshot AI directly entered the office intelligence agent Track (sector). Feishu's window to break through growth ceilings with standalone SaaS was narrowing.

ByteDance had repositioned Feishu within its AI strategy. At the January 2026 all-hands meeting, Liang Rubo judged AI to represent "at least a PC+Web-scale peak." At the June 23 FORCE conference, he made it clearer: "Scaling the AI peak" was ByteDance's top priority, with Volcano Engine's MaaS becoming a foundational business. 36Kr summarized this widely: "Everything but AI is contracting."

Contraction coexisted with investment. Public reports indicated ByteDance increased AI new business investment in 2025; 2026 capital expenditures rose to over 200 billion yuan, with ~85 billion earmarked for AI chip procurement, and discussions to raise that ceiling further. Such investment demanded a sufficiently large monetization outlet.

04

What Doubao Lacked, Feishu Had

The preferred outlet was initially Doubao, but its financials were unimpressive.

By user scale, Doubao dominated: MAUs grew from 345 million in Q1 2026 to 382 million in June, with DAUs exceeding 200 million and daily token usage surpassing 180 trillion. Revenue told a different story—according to LatePost, Doubao's daily revenue was under 1 million yuan, primarily from in-app e-commerce commissions. Extrapolating from Volcano Engine's public API pricing, Doubao's daily inference computing costs likely reached tens of millions of yuan. Massive users brought not commercial value but sustained bleeding.

In May, news of paid messaging emerged, and Doubao's MAUs dropped by 6.1 million that month—its first significant decline since launching in 2023. In June, despite controversy, Doubao formally launched paid subscriptions. More challenging was user demographics: Citigroup research noted that students and elderly users dominated Doubao's 300+ million MAUs, with scenarios concentrated on casual chat and information lookup, averaging under one hour of daily use. Working professionals and developers with high-frequency productivity needs found it inadequate, preferring ChatGPT, Claude, or domestic alternatives like Kimi.

Doubao wasn't short on users but lacked enterprise scenarios and high-paying demographics. Feishu had both: its enterprise intelligence platform Aily, multidimensional tables with over 10 million MAUs, and years of accumulated enterprise workflows and more mature paying clients. The Doubao Enterprise Edition, co-developed with Feishu's team, began internal testing among select Feishu clients in late June—employees could summon Doubao directly from Feishu's knowledge Q&A, documents, group chats, or meeting interfaces. By mid-July, Feishu's "knowledge Q&A" feature quietly switched to Doubao.

This merger wasn't Feishu being abandoned but a mutual need: Doubao gained enterprise access and permissions, while Feishu tapped into large model technical capabilities and Volcano Engine's sales channels. A person close to ByteDance put it bluntly: "A business with tens of billions in ARR that can't become an entry point can only unleash greater potential by merging with Doubao and Volcano."

05

The Entry Point Moved

Zooming out, Feishu's restructuring was part of a broader industry trend that summer.

On July 20, Tencent integrated parts of the QClaw product center into WorkBuddy, which it had already designated a "super project" overseen personally by Pony Ma. On July 27, Alibaba consolidated multiple Agent products into "Qianwen Office," handed to DingTalk's new CEO with pricing directly targeting WorkBuddy. Huawei took a more radical approach: Xiaoyi, its AI assistant since inception, was becoming the unified AI entry point across all scenarios.

Within ten days, major tech firms had all consolidated their AI office strategies. Behind this lay a set of numbers: iResearch reported that in June, 17 mainstream desktop AI office agents in China saw over 60 million total visits, with Tencent WorkBuddy leading at 20.97 million—exceeding the sum of second and third place. ByteDance's Trae ranked second with 12.79 million visits. These visits occurred on PCs, far more valuable than mobile fragmented time—users activated Agents for writing long documents, handling complex spreadsheets, and managing corporate emails, reflecting clear productivity demands. Enterprise users weren't chatting; they were working.

A clear judgment emerged: the first work software employees open won't be collaboration tools but AI assistants, with office functions extending AI capabilities. Since late March, DingTalk, Feishu, and WeCom have all open-sourced their CLI interfaces—the logic is clear: when Agents become the first window opened at work, traditional office software must retreat from entry points to become execution engines. Microsoft 365 Copilot already had over 20 million paid enterprise seats, adding 5 million in a single quarter. IDC predicted that China's enterprise-grade token consumption would grow ~20x from 2025 to 2026.

The entry point had moved—this was the true context of Feishu's restructuring. Among office software in its golden decade, Feishu was the most advanced. But when "advanced" was redefined by AI, clinging to an independent entry point became the riskiest choice. Maintaining separate teams would only drain resources, raise client cognitive costs, and even create internal competition. Unifying products, sales, and services was essential for enterprise AI to deliver real value.

06

What Those Dots Connected To

Returning to Xin Xie's speech title.

When Feishu launched in 2016, it aimed to solve how information flows within organizations. A decade later, enterprises' top pain point became: Can AI read corporate knowledge, understand organizational permissions, invoke business tools, and ultimately get work done? Feishu's decade-long accumulations didn't lose value—quite the opposite. Corporate documents, messages, organizational relationships, permissions, and business data were prerequisites for Agents to enter organizations. Their value just wouldn't be realized through "a standalone office software" but as ByteDance's AI foundation.

For ByteDance, this was arithmetic: facing 200 billion yuan in AI infrastructure investment, Feishu with 3 billion yuan in revenue couldn't retain independent destiny-defining power. For the industry, it signaled a shift: the first half focused on concepts and parameters; the second half would see which solutions truly solved problems, with AI's center of gravity moving from "internal competition" to "entry point contention." For Xin Xie personally, in ten years he had grown a ten-person team into a 4,000-person independent BU, only to willingly place it onto a larger platform.

The dots he connected in his speech now led clearly: from HR to tools, from tools to Feishu, from Feishu to AI. These dots formed not Feishu's empire but a cornerstone of ByteDance's AI ecosystem.

This isn't a failure narrative. Feishu remains a strong product with loyal clients and maintains double-digit growth. But in the AI era, the definition of "strong product" is being rewritten—advanced products may not own their era, but they can choose to remain part of it.