Summary of Key Points
ByteDance has undergone significant integrations of its AI business within a month: first, it merged the Lark product team with DouBao, and assigned Lark’s marketing and sales services team to Huo Shan Engine; then, it integrated the TRAE AI programming team and the Kou Zi intelligent agent development platform into the DouBao system. The primary goal of these actions is to make DouBao the “backbone” of the company’s AI business. By leveraging DouBao’s existing C-end user base, combining Lark’s enterprise scenarios, TRAE’s programming capabilities, and Kou Zi’s intelligent agent technology, the company aims to focus resources on enhancing AI productivity in tasks such as office work and programming. Huo Shan Engine will be responsible for sales and services in the enterprise market, ultimately forming an AI-to-B (AI for businesses) strategy where DouBao handles product development and Huo Shan Engine handles commercialization.
Detailed Analysis
1. Why Did DouBao Suddenly Become the “Leader” of the AI Business?
DouBao’s rise to the forefront is due to two main advantages: first, it has a solid C-end user base (many ordinary people have used DouBao for chatting); second, the company wants to transform AI from a chat tool into a tool for practical work. Previously, DouBao was mainly used for casual conversations, but now ByteDance intends to integrate it into high-frequency work scenarios such as document writing, code editing, and automatic meeting summary generation. To achieve this, DouBao needs the capabilities of other teams: Lark’s enterprise customers and collaborative office features, TRAE’s AI programming skills, and Kou Zi’s intelligent agent technology. By bringing these teams under one umbrella, DouBao can evolve from a general chat assistant to an AI platform that can solve real work problems.
2. Lark’s Reorganization Doesn’t Mean It’s Being Abandoned; Instead, It’s Getting a New Role
Lark was once an independent office product that competed with DingTalk and WeCom for market share. However, with AI becoming a core strategy for ByteDance, Lark’s value has changed. It is no longer just an office software competing with rivals; rather, it serves as a “bridge” for DouBao to enter enterprises.
- The Lark product team was merged into DouBao because Lark has mature office features (documents, approvals, collaboration), and DouBao’s AI functions can be directly integrated into these scenarios. This allows enterprise users to use Lark and, in turn, promote DouBao’s adoption within companies.
- Lark’s marketing and sales team (GTM) was transferred to Huo Shan Engine, which specializes in providing enterprise services. Lark’s sales team has accumulated a large number of enterprise customer relationships, and by integrating this team with Huo Shan Engine, the “Lark + DouBao AI” combination can be more effectively sold to businesses. In short, Lark has transformed from an independent product to a platform that supports DouBao’s AI strategy.
3. The Integration of TRAE and Kou Zi: Avoiding Internal Competition
TRAE focuses on AI programming (helping programmers with coding, bug detection), and Kou Zi develops intelligent agent platforms. Although these teams were working independently, ByteDance realized that their capabilities align well with DouBao’s focus on AI productivity. Continuing to operate them separately would lead to redundant development and internal competition, wasting resources. By integrating them into DouBao, their technologies and experiences can be shared: TRAE’s programming tools will become part of DouBao’s product line, and Kou Zi’s intelligent agent capabilities can be integrated to create customized AI assistants for enterprises. This not only reduces internal friction but also enhances DouBao’s functionality.
4. ByteDance’s Organizational Adaptability: “Pulling the Reins After the Race”
ByteDance’s approach is to allow multiple teams to experiment simultaneously when the direction is unclear (referred to as “horse racing”), and then immediately focus resources on the winning team once a clear direction is established. This integration of AI teams is a typical example of this strategy:
- At the beginning of the year, DouBao, Lark, TRAE, and Kou Zi were each working on AI productivity-related projects (DouBao on general assistants, Lark on office AI, TRAE on programming AI).
- Mid-year, ByteDance determined that AI productivity was the right direction, and DouBao was the best candidate as the core platform, so the other teams were integrated into DouBao to stop duplicating efforts.
This approach allows for testing different options early on and then concentrating resources on the most promising direction. For example, Toutiao was once a core business but was merged into Douyin when short videos became more popular. ByteDance never retains a business solely because it was once important; it focuses on whether the business is still relevant to its current strategy.
5. The Advantages and Risks of This Adjustment
Advantages:
- Resource Consolidation: By bringing together dispersed teams and capabilities, ByteDance avoids duplication and accelerates the development of DouBao as the core AI product.
- Utilization of Existing Resources: Lark’s enterprise customers, TRAE’s programming skills, and Kou Zi’s intelligent agent technology can be leveraged without starting from scratch.
- More Efficient Commercialization: Huo Shan Engine handles sales, and DouBao focuses on product development, allowing for a more efficient rollout of AI products to enterprises.
Risks:
- Team Stability: Frequent organizational changes may cause employees to lose confidence in the long-term success of their efforts.
- Strategic Mistakes: If ByteDance misjudges the direction of AI productivity, the resources invested in DouBao could be wasted.
- Integration Challenges: Different teams may have different cultures and technical backgrounds, and integrating them can be challenging and potentially lead to issues.
In summary, ByteDance’s move is a major bet on AI productivity, aiming to make DouBao the company’s next core business, just like Douyin. Whether this strategy will succeed depends on the quality of the integrated products and the acceptance among enterprise customers.