Summary of Key Points
Recently, major internet companies have engaged in two types of reflection on the implementation of AI: Meituan has reflected on the waste of computing power caused by its AI promotion efforts (daily costs in the tens of millions, with flawed models interfering with business operations), while Tencent nearly abandoned a promising AI project due to cost-cutting measures. These seemingly opposite reflections point to the same core issue: how to truly transform the AI revolution into productive forces and commercialize its benefits. The companies have ended their internal competition and are now competing on multiple fronts in the AI office space, while also navigating the complex relationship with paying users, particularly corporate clients. The ultimate success will depend on their ability to find a balance that benefits both parties.
Detailed Analysis
1. Major Companies' Reflections: The Dilemma of AI Implementation
After the surge in popularity of AI, these companies have either gone too far or been too cautious, both resulting in setbacks:
- Meituan's Waste Dilemma: By attempting to promote AI across all its operations, Meituan incurred daily costs of 10 million yuan, and the incorrect information produced by its models disrupted its business operations (for example, recommending wrong strategies to merchants). This indicates that although there is an abundance of AI resources, few companies have been able to generate measurable benefits.
- Tencent's Cost-Saving Dilemma: The AI project CodeBuddy (later renamed WorkBuddy) was nearly canceled due to cost-cutting efforts, but fortunately, it was kept. This reflects the extreme tendencies within large companies when evaluating AI projects—either blindly investing money or giving up due to a lack of immediate returns.
Both reflections essentially ask the same question: how can AI be used effectively to avoid wasting resources and truly contribute to business success?
2. Internal Integration: Ending Internal Competition to Capture Core Resources
Companies are merging their AI office product lines. For example, Tencent has combined QClaw and WorkBuddy, ByteDance has integrated Feishu, DouBao, and KouZi into "DouBao Work," and Alibaba has integrated three lines to launch QianWen Office. Why?
- Ending Internal Competition: Previously, different departments were developing their own AI projects, leading to the dispersion of resources. Now, they are consolidating their efforts to focus on a common goal.
- Competing for Scarce Resources: The most valuable assets in the AI era are user access and corporate data. Whoever controls the office platforms will gain access to essential business data, which, in turn, allows for better model training and attracts more users, creating a positive cycle ("flywheel effect"). However, AI is still a capital-intensive field: Tencent's capital expenditure increased by 176% in the first half of the year (mainly for purchasing AI computing power), and Alibaba's AI lab incurred a loss of 13.8 billion yuan. Why continue investing? Because they fear falling behind; if other companies invest but not they do, shareholders may doubt their commitment, potentially leading to a drop in stock prices (this is known as the "prisoner's dilemma"—investing is always the best choice, regardless of what others do).
3. Competitive Strategies: Big and Small Players, Each with Their Own Strengths
The competition in the AI office market is not just about individual products but encompasses infrastructure, platforms, and applications. Each player is leveraging their unique strengths:
- Large Company Camp:
- Tencent WorkBuddy: Builds on its ecosystem, integrating with Tencent Documents, cloud storage, and Lexiang, allowing for local file management, remote computer control, and coding.
- ByteDance DouBao Work: Closely integrated with Feishu, enabling users to access chat records, documents, and meeting minutes without re-entering information.
- Alibaba QianWen Office: Supports desktop, cloud, and enterprise collaboration, and integrates with DingTalk, offering features like automatic meeting summaries, document creation, and scheduling.
- Baidu ZaiZi: Leverages search and data management capabilities to target the corporate payment market.
- Smaller Players:
- Kingsoft WPS Comate: Uses its document management strengths to organize corporate knowledge, making the AI more intelligent with increased user usage; however, the migration cost is high (all data must be re-imported if switching to another product).
- SenseTime AI Box: Sells hardware devices that provide localized AI tools, addressing concerns about data security.
- JieYueXingChen: Develops AI-powered smartphones without apps, using an AI operating system to connect various applications, with the goal of solving work-related problems.
This competition is a long-term battle that tests both product quality and financial resources.
4. The User as the Final Judge: Turning AI from a Cost Driver to a Profit Generator
Regardless of the competition, the ultimate decision-makers are the paying users, especially corporate clients. The battle revolves around two main aspects:
- Explicit Pricing: Current AI office pricing is chaotic, with subscription models, pay-per-use models, and free trials. The issue is that the same amount of tokens may have vastly different values depending on the use case (e.g., generating simple text vs. complex business plans). Users prefer payment based on the actual benefits AI provides. The future trend is towards pay-per-effect models (e.g., AI marketing tools charging a percentage of sales). Gartner predicts that 70% of AI services will use this model by 2027.
- Implicit Service Delivery: Corporate clients face three major challenges: data silos, non-standard processes, and employee resistance to new tools. Suppliers must help address these issues (e.g., by integrating data, customizing processes, and training employees) before AI can be effectively utilized. It's not just about selling a product; it's about providing comprehensive support to make the technology useful.
In the end, the winner will be the one that finds a balance that satisfies users and generates profitable revenue.
Conclusion
The competition in the AI office market is just beginning. From internal integration to external competition and user adaptation, every step presents challenges. Whether AI can transform from a cost center to a profit generator depends on its ability to solve real business problems and provide value to users. For individual users, more intelligent office tools may emerge in the future; for companies, however, the choice of AI products will require a more rational approach—focusing not just on technology but on its potential to generate revenue.