Summary of Key Points
After CEO Li Zhifei used AI to quickly develop a prototype of the “Feishu” (a communication tool) version, he noticed that while individual efficiency improved due to AI, team collaboration remained traditional, resulting in overall organizational inefficiency. This led to what he called the “ghost efficiency” trap—where the stronger individuals made the system even more congested. The company reduced its losses significantly by cutting back on businesses and scaling down, but this also caused a decline in revenue. They are now exploring how to implement AI-driven organizational changes (with the goal of collaborating with 100 companies by the first half of 2026) and shared their entrepreneurial insight that choosing the right business direction is more important than chasing trends.
1. AI Enhances Individuals, but Teams Are Still Stuck in the “Old World”
Li Zhifei used AI to write hundreds of thousands of lines of code in just three days to create the Feishu prototype and felt that the old ways were no longer worth sticking to. However, during a team meeting, he realized that everyone was still working on a weekly/monthly schedule, with front-end and back-end tasks being handled separately, just as before. This contrast caused him anxiety: even if everyone used AI, organizational efficiency might not increase tenfold because the rapid progress of individuals could outpace the ability of teams to collaborate effectively.
2. Why Do Super Individuals Hold Back Organizations? The Problem with “Ghost Efficiency”
Li Zhifei made an analogy: a super individual is like a “super spectator” who enjoys directing AI tasks without the need for meetings; but an organization is like a team that requires collaborative effort from multiple people. If individuals prefer to work independently and avoid communication, the organization will become inefficient. He proposed a formula: Super Individual Ability × Collaboration Mechanism = Super Organization (multiplication, not addition). No matter how strong an individual is, if the collaboration mechanism is weak, the result will be zero. For example, using different tools for tasks (like Feishu/DingTalk for meetings and Figma/Cursor for work) leads to wasted time and low efficiency.
3. Cutting Back on Businesses and Scaling Down: Less Losses, but Lower Revenue?
Over the past two years, “Outdoor Ask” reduced its staff from 1,200 to 150 and eliminated its car-related B2B business (since licensing each car cost only 7 yuan, generating 7 million yuan from 1 million cars but requiring 200 employees, which was not cost-effective). They also dropped some old businesses that seemed promising in the AI era. The 2025 financial report showed a 17.2% decrease in revenue to 323 million yuan, but losses were reduced by 90.5% to 68.9 million yuan. Although the reduction in costs was successful, new business revenues have not kept up. Therefore, they are focusing on collaborating with companies to develop AI-driven products.
4. Challenges in Transforming Organizations with AI: Needing More Practice with 100 Companies
Turning their internal AI tools into general-purpose products presents significant challenges due to differences in employee habits, company cultures, and leadership styles. Some CEOs are willing to try, but there is resistance within the companies. Li Zhifei believes that without change, companies will fail and aims to collaborate with 100 companies by the first half of 2026 to make the product more adaptable to various organizations.
5. An AI Veteran’s Entrepreneurial Advice: Choose a Business Direction With Three Criteria
With 14 years of experience in entrepreneurship, Li Zhifei has tried voice assistants, AI hardware, and large models but hasn’t found a scalable business model. He advises young entrepreneurs to choose a direction that is important, exciting, and promising for the future. His advice is to select a business direction that meets three criteria: something you enjoy, something you can execute well, and something society needs. Avoid chasing trends or getting bogged down in details; maintain passion for innovation to stay competitive.
This news article reflects the challenges of organizational transformation in the AI era: while individual efficiency can be improved easily, upgrading team collaboration is more difficult. Companies need to find a balance between cutting outdated businesses and exploring new collaborative models. Li Zhifei’s experience illustrates the reality faced by many AI companies—technology is advancing rapidly, but organizations and people are struggling to keep up.