Summary of Key Points
In an interview, Kai-Fu Lee made four key observations regarding the AI transformation of businesses:
1. AI has evolved from a “consultant that provides advice” to an “executor that can put plans into action,” directly impacting the software industry—standard software is becoming less valuable, while custom software is gaining in value as AI makes it more affordable.
2. Three common concerns CEOs have when initiating AI transformation can be addressed with the use of “CEO-level AI tools”: stop worrying about which model to choose, map out the company’s structure, and activate dormant data.
3. AI transformation should be a top-priority project; relying solely on the CIO will limit the impact to low-value applications. CEOs need to restructure their businesses at a strategic level.
4. The future organizational unit of businesses will consist of a “DRI (Directly Responsible Individual)” along with a group of AI Agents. One person, with the help of AI, can manage an entire business line, breaking through the limitations of traditional human-based organizations.
1. From “Talking” to “Doing”: Is Standard Software on the Decline?
In the past, AI could only provide suggestions through conversation. Now, with enhanced coding capabilities (the ability to write programs and automate processes), it has become an actual executor. For example, AI can prepare reports for board meetings by organizing data, predicting issues, and creating PowerPoint presentations—all in a closed-loop process. The impact on the software industry is:
- Value of standard software (such as Excel, Oracle) is declining: Companies used these tools not because they were perfectly tailored to their needs, but because custom software was too expensive (requiring thousands of engineers). Now, with AI, companies can adapt their software to their requirements.
- Value of custom software is increasing: AI can quickly understand a company’s business model and data assets, allowing teams of just a dozen people to create customized systems that were previously only accessible to large firms. Traditional businesses are no longer constrained by standard software and can finally make it fit their needs.
2. CEO’s AI Concerns? Three Solutions with “CEO-Level AI”
Many CEOs struggle when implementing AI transformation: they don’t know how to proceed, fear that their efforts will be ineffective, or worry about being outperformed by competitors. Kai-Fu Lee offers three solutions:
1. Stop worrying about the model: The top models available in China are similar; having just a model is not enough (just like having an Intel chip 40 years ago was useless without an operating system). The key is to make sure AI understands your company’s specific needs.
2. Map out the company’s structure: Clearly define the business architecture and decision-making processes, and indicate which data is truly important to avoid confusion (e.g., mixing financial reports from Europe and Asia).
3. Activate dormant human-generated data: The most valuable data in a company often comes from meetings, customer complaints, and sales communications—these contain strategic insights and execution details. By making this data accessible to AI, companies can gain a better overall understanding.
3. Don’t Let the CIO Lead the AI Transformation!
The role of the CIO is to ensure system stability, with a focus on tools and low-risk deployments. If the CIO takes charge of AI transformation, they might only implement low-value tasks like customer service responses and meeting summaries, missing out on the revolutionary potential of AI.
True transformation requires the CEO to take the lead, thinking strategically about how AI can help increase revenue, boost profits, and speed up product release times. Companies that transform quickly (such as those in the Fortune 500) place AI at the core of their business operations, not just using it as a tool for IT departments.
4. Will Future Businesses No Longer Need Thousands of Employees? Rebuilding with DRI + AI Agents
In the past, companies grew by hiring more staff, but larger teams led to increased management complexity (e.g., a VP managing thousands of employees). The future organizational structure will be:
- DRI (Directly Responsible Individual): Responsible for setting goals, making critical decisions, and conducting business negotiations (tasks where humans excel).
- AI Agent cluster: Handles execution, coordination, and process follow-up (repetitive tasks where AI is more efficient).
For example, one DRI, with the help of a few AI Agents, can manage an entire business line. This isn’t about reducing staff; it’s about freeing up resources for more valuable activities, allowing smaller teams to accomplish what used to require larger companies.
Conclusion
Kai-Fu Lee’s main point is that AI has moved from a technical concept to a practical business tool. To reap the benefits of AI, businesses must have CEOs take the lead in using it to transform their operations and structures. The future is not about AI replacing humans but about combining human capabilities with AI to create even greater value.