It’s been cold in Beijing these days, but Baidu World Conference 2026 is particularly hot.
The content on site was super rich, and Baidu is still the clear leader in AI. But what impressed me the most was the interview video at the beginning. Baidu’s digital person Xi Jiajia replaced the real person in the interview with Robin and asked many in-depth questions.
This is also the world’s first digital human interview video. It allows digital humans to interview Robin and AI to play mahjong with Robin. Baidu is really good at playing.But I think what’s more worth talking about is the question Xijia asked Robin in the video: If AI becomes more and more powerful, can digital people replace real people as CEO?
This sounds a bit surreal, but I think this is why Robin Li wants to focus on “intelligent organization” this time.
In the past two years, many companies have been promoting AI transformation. Employees use AI to write plans and do analysis, and partners from Lei Technology have even tried to use intelligence to report on the IFA exhibition. AI is indeed very useful, but the question is: as personal efficiency increases, does the overall company efficiency necessarily become higher?
Not really sure. For example, an employee uses AI for half an hour to make a plan that originally took a day to complete, but the process still needs to be followed, and cross-department review and confirmation still takes a day. Microsoft calls this situation the “Enterprise AI Transformation Paradox.” The Boston Consulting Group has also proposed that companies must redesign the company for AI rather than letting AI adapt to the original company.
As we all know, someone in the industry has already proposed the “five-layer cake” theory, which talks about how AI capabilities are built. But AI technology is just a tool, and making good use of this tool is the focus of enterprises. Without changing the original division of labor and processes, it will be difficult for enterprises to make good use of AI technology and transform AI capabilities into organizational productivity.
How has the business changed? Baidu still has a say in this matter. After all, it has been doing one thing over the years: AI industrialization.
At the 2025 Baidu World Conference, Robin Li proposed that enterprises should internalize AI capabilities; at this year’s Create conference, he shared his thoughts on the self-evolution of enterprises; at this time at Baidu World 2026, his thinking went further towards the organization itself. The “intelligent organization” he proposed is actually the new factory building in the AI era. Companies must build working environments such as factories and assembly lines so that AI can be integrated into real business.
To put it bluntly, personal efficiency improvement is peripheral work. Only by building an intelligent organization can AI enter the core of the enterprise, drive growth from the inside out, and release organizational productivity.
So, how to build an intelligent organization? Robin Li put forward three principles:
In the past, companies assigned work according to positions. In the future, work can be divided based on tasks, allowing employees and agents to complete the work together.
In the past, the experience accumulated by employees was kept in their own minds, and everything relied on the “teaching and guidance” of their masters. Now, the experience can be accumulated and can be used by other employees and even agents to transform personal experience into organizational capabilities. Finally, when using AI, you must be able to evaluate results, identify problems, and then use this feedback for the next task. “Measurable and superimposed looping allows enterprise organizations to continuously evolve the process. This is the “PDCA” in the AI era.
In fact, Baidu is not alone in exploring intelligent organizations. McKinsey proposed Agentic Organization and Microsoft proposed Frontier Firm.
Everyone uses different terms, but the conclusion is the same: AI is getting stronger and stronger, AI is breaking through the boundaries of traditional jobs, and companies must reorganize people and AI.
Xijaga: What you’re looking at is that the organization is not going away, but it’s trial and error faster and iteration is denser. What is the core of the last remaining part of the organization?
Robin: This is a good question. The core thing is actually people.An organization must have a mission at its core. Why it exists is defined by humans. This is the biggest difference between humans and AI. Humans have a mission, but as for AI, I don’t think it does.
This actually answers the question at the beginning: Can AI replace a real person as CEO? AI can take on more and more tasks, and can even manage teams and manage projects, but the “vision, mission, and values” of an organization must be defined by people.
Therefore, the criterion for judging the success of enterprise AI transformation is not how many tokens are burned, how many AI tools are introduced, or how many agents are deployed.It’s about whether AI capabilities can be integrated into business, so that personal efficiency can be transformed into organizational capabilities, and the execution of tasks can accumulate into the organization’s ability to continue to evolve, and even become a moat.
How to build a good intelligent organization will be a must-answer question for every enterprise in the AI era.
Related Reading
- “Bring the boss back”, after Huang Renxun resigned, Nvidia employees couldn’t sit still for the first time2026-11-15
- At the age of 29, she became the core figure of Bit Group AI2026-11-15
- Byte’s 2029 and Tencent’s “Half People”2026-11-15
- Jen-Hsun Huang is out of class, and there is no permanent master in Silicon Valley2026-11-14
- Huang Renxun resigns with “double standards”, Musk: Welcome to Tesla2026-11-14