Chen Chunhua's "Management in the AI Era": The four fundamentals of planning, organizing, leading, and controlling must be redone

Source: Notemans

Many entrepreneurs have talked to me about this. They feel that they've spent the money, taken the courses, and installed the tools, yet the company is still the same as ever.

Where's the problem?

I went through the whole book and picked out the few things most useful to you. I suggest you read to the end.

What did these systems do? Nothing more than move your original management actions online. Approvals are still approvals, reports are still reports, and the departmental walls are still those same walls.

What did traditional management stand on? Information scarcity.

Bosses knew more than employees, so there had to be hierarchy; information gets distorted as it passes down layer by layer, so processes had to be set; it was hard to measure how much each person did, so there had to be KPIs.

The book gives an example: at a large retail enterprise, when a store submits a request to transfer stock, AI uses historical sales, inventory turnover, and supply chain conditions to calculate the optimal allocation plan on the spot and generate an execution order on the management platform. Managers don't need to intervene manually, but they can see resource flows, task progress, and exception alerts at any time.

Management actions themselves become data, and the management platform merges with the data platform into one.

Since the old premise is gone, the practices have to change with it.

Chen Chunhua proposed a concept: the water-like organization.

What does that mean?

A traditional company is like a solid: people and resources are nailed to their own positions. To get something done across departments, you file a process, hold meetings, get signatures—it takes a long detour to accomplish anything.

Information, resources, and talent should follow the task, not the department.

AI counts as a formal member of the organization, which Chen Chunhua calls a "digital employee." When a R&D team is testing a new product, the digital employee doesn't wait for your instructions; it watches market changes, small supply chain fluctuations, and anomalies in user behavior on its own, warning of risks in advance and sometimes spotting opportunities earlier than people do.

Management used to be about harnessing resources; from now on, it's about designing how people and machines work together.

ByteDance is often described as a company that looks like an App, meaning every action and every decision flows through the system, traceable and optimizable.

So what do managers do at this point? Set the context: where resources flow, who participates, who verifies first—the system helps with scheduling.

The whole process of an idea from birth to completion is also recorded by the system. Who proposed the core improvement, who ran the experimental validation, who led the team through iterations—every action has data. Whose contribution is greater can be seen at a glance.

Traditional assessment settles the accounts at year-end; Haier keeps a running ledger every minute of every day.

The two cases are really saying the same thing: from now on, a company's competitiveness won't depend on how many people you have, but on whether your system can recognize each person's contribution in real time and amplify it in real time. III. All four management fundamentals must be redone

Management has been taught for over a hundred years, and its fundamentals are just four: planning, organizing, leading, and controlling.

Planning has gone from a PowerPoint slide to a real-time navigation that keeps calculating based on the latest data and constantly offers the best route for the moment.

The book has a practical tip: when instructing AI, stop using vague phrases like "please optimize logistics costs." Try stating the conditions clearly: on the premise of ensuring that the 24-hour delivery rate for customers in North China is no lower than 98%, adjust the proportion of forward warehouses and find the best balance between logistics costs and inventory costs.

Your level at giving instructions is your level at planning.

The book also describes a tool: the digital twin decision cockpit. It creates a real-time synchronized digital mirror of the real world. For example, if a cargo ship in the supply chain is delayed, the virtual route in the decision cockpit immediately turns red, automatically popping up the number of affected orders, how much compensation is owed, and what alternatives exist.

In the past, the fundamental of management was after-action review—looking back once things had passed. Now the fundamental has become "sensing fluctuations": before things happen, the system is already running simulations ahead of time.

Second, organizing.

What do you do when collaboration gets stuck? The old way is to talk to people. The new way is to check the rules: is the division of labor unclear, or is feedback too slow? Straighten out the rules, and the organization repairs itself.

Managers' identity has changed too. They used to be the people who mediated interpersonal conflicts; from now on, they are the people who design the rules of collaboration.

Third, leading.

Algorithms only recognize probability and efficiency; people need a sense of security and emotional connection. Run a company entirely by algorithm, and it will become more and more devoid of human warmth.

An algorithm can calculate that laying off 100 people is the optimal solution, but how to properly resettle those 100 people is a decision people have to make.

Fourth, controlling.

Traditional control is watching KPIs, which in essence is guarding against people making mistakes. In the AI era it's the reverse: the task of control becomes inspiring people.

Chen Chunhua put it brilliantly: watching metrics is traditional management; watching "logical increments" is what counts as intelligent management.

What is a logical increment? It's what new thing a person contributes to the company—a new idea, a new way of doing things, all of it counts. The system must recognize people's creative process and actual contributions, not just look at the dry final number. IV. Jobs are disappearing, roles are taking the stage

Permissions follow roles and are dynamically allocated with tasks. When the task ends, permissions are revoked.

When roles change, the way money is paid has to change too.

How was money paid before? By job grade, by seniority—the longer you stayed, the more you got.

Super-individuals don't buy that. What they want is to be paid by role contribution: the more value you create, the bigger your share.

Incentives have changed direction too. They used to rely on the carrot and the stick, pulling people along from the outside.

From now on it's about ignition: making a person want to do something themselves, and what they want to do happens to be exactly what the company needs.

The book has a case: Studio F was originally a standard game development team of 200 people. To get those 200 people to work together, functions were sliced up finely—art team, design team, programming team, numerical design team, marketing team—a dozen or so small groups each handling one piece.

What happened after AI was introduced? The work of 200 people was taken over by 5 super-individuals.

What does this show? The advantage of having many people is depreciating, while individuals who can use leverage are appreciating.

Chen Chunhua says: any work that can be written as logic, repeated, and has a standard answer will no longer command a premium.

So what commands a premium? Three places.

First, asking questions.

Second, aesthetics.

In the book's example, Lao Zhou, creative director of Liuguang Intelligent Design Studio, used to be the number-one quality inspector, spending every day going in circles revising drafts. After adopting AI, he stopped worrying about how designers write prompts and put his energy into product selection and aesthetics. AI produces a massive number of options at once, and drawing on years of industry accumulation, he picks the few that best fit the brand's character and move people most.

Third, giving meaning.

Algorithms can pile up words, but they cannot understand the suffering and joy of the people behind the words.

What will be valuable in the future is cross-disciplinary insight, judgment of right and wrong in complex situations, and the ability to raise key questions.

Relying solely on old knowledge in your head won't work anymore. Six, how do traditional enterprises transform?

After all this, here comes what traditional enterprise bosses care about most. My company doesn't have ByteDance's foundation—what should I do?

First jump: digitalization. First move the business online so that everything can be recorded.

Second jump: datafication. Let data flow between processes and break down information silos one by one.

Third jump: intelligentization. Humans and machines work together, and intelligence becomes the company's everyday capability.

Digitalization is the means, managing connections; datafication is the process, managing collaboration; intelligentization is the result, managing symbiotic value.

What does it look like after the third jump is complete? The book has an unnamed Company S that illustrates the point well.

AI agents monitor popular elements on global social media 24 hours a day. The moment a hot spot is spotted, a visual agent immediately produces sketches, a copywriting agent simultaneously generates descriptions in 12 languages, a logistics agent works out the warehousing and distribution plan, and the instructions are pushed directly to the digitalized workshop for second-level prototyping.

Previously it took weeks for a new product to go from spotting a trend to prototyping; now it's at the second level.

Let's look at a few more cases in manufacturing and services.

Midea's Nansha factory uses reinforcement learning for production scheduling. When AI discovers that a certain part is delayed in transport, it reschedules the entire plant's production within milliseconds, so it no longer stops and waits for materials. Machine vision is used for quality inspection, cutting the missed-detection rate from 5% to below 0.1%.

By 2026, Midea's Jingzhou washing machine factory obtained the world's first certification for a multi-scenario agent factory, with 14 agents managing 38 core production scenarios. In 2025 alone, Midea saved more than 15 million work hours and 700 million yuan in costs through AI.

Feihe, a milk powder company, took digitalization all the way to the cows. Each cow wears a smart collar, and AI collects motion data dozens of times per second, distinguishing rumination, feeding, movement, and sleep. A cow ruminating 450 to 550 minutes a day is considered normal; once rumination drops by more than 15%, the system notifies the veterinarian one or two days before the illness onset.

Laoxiangji, a restaurant company, took another path.

Chinese fast food has a curse: once standardized, it loses its "wok hei." Laoxiangji uses algorithms to break down and study the movements of tens of thousands of chefs, assisting chefs to ensure that every freshly stir-fried dish at the lunch peak still has the taste of a home kitchen.

The previous six points are all about organization and business. The last one is about you yourself.

First, digital literacy, which addresses vision.

Second, moral judgment, which addresses the standard.

Third, psychological coaching, which addresses motivation.

The super-individual's greatest enemy is a sense of meaninglessness. When AI can do 90% of the work, people easily doubt the value of their existence. Managers must be able to see the person behind the metrics and give them a sense of meaning.

For veteran managers, this is a process of power being taken away. Where did their sense of achievement come from before? From control. Subordinates had to ask for instructions on where to go, plans needed your signature, meetings waited for you to make the call. AI puts the entire decision-making process out in the open, and many managers suddenly feel useless, panicking inside.

It's like an old driver used to a manual transmission sitting in a fully autonomous driving seat, hands hovering in the air, always wanting to grab back the steering wheel.

Thus comes "defensive management": claiming it's risk control while actually running small-circle decision-making, finding fault with algorithms, and finding ways to take power back.

If this hurdle isn't cleared, all the methods discussed above cannot be implemented. Conclusion

Planning, organizing, positions, and control all have to change accordingly, and so do managers themselves.

(End of article)

www.smartcity.team

Disclaimer: Please indicate the source as "Smart City Industry Analysis (www.smartcity.team)" and the relevant author when reprinting. Except for original content, some information on this site comes from the Internet. If the relevant content involves infringement of rights, please contact the contact information at the bottom to delete it.

[Digital Technology Special Research] Community

In-depth peer exchange, focused on smart city and digital technology industry research and material accumulation.

Continuously updating the latest industry white papers, industry research, policy developments, and implementation solutions to help industry practitioners efficiently accumulate knowledge and share experience.

Massive essential content continuously iterated, providing frontier vision, connecting industry talent, carrying out case co-creation and resource exchange, and jointly building a high-quality industry exchange atmosphere.

#post-##

Recommended for you

CAICT releases the top ten keywords of the 2026 AI industry: unfolding from three levels—technology, application, and ecosystem—outlining a complete picture of AI's accelerated diffusion in 2026 (with CAICT Vice President Wei Liang's interpretation PPT and the report text)

Meng Qingguo: Six major changes in data element circulation governance in the intelligent era; the key to governance is to promote the transformation of data into callable, governable, measurable, and profit-shareable intelligent productivity

Wang Chunhui: "AI+" reshapes the development logic of new quality productive forces

Ministry of Finance "Local Government Special Bond Project Implementation Plan Template (Trial)": focuses on adding a chapter on ex-ante performance evaluation; special bonds are shifting from "emphasizing issuance scale" to "emphasizing project quality and full-cycle performance"

Ouyang Rihui, member of the National Data Expert Advisory Committee: Scenario innovation drives digital talent cultivation

In-depth report: Full interpretation courseware PPT of the "Opinions of the CPC Central Committee and the State Council on Developing New Quality Productive Forces" (3 editable versions)