Xie Hong/Wen

Recently, the author conducted research at a small home appliance company in Zhongshan, and the company's head, Mr. Yang, shared an interesting change: starting this year, they have partnered with the Guangdong Small and Medium sized Enterprise Development Promotion Association to establish an "AI Innovation Laboratory" and promote AI tools within the company. At first, the company thought that AI transformation, like traditional informatization, was led by the IT department - building platforms, building systems, and conducting training. But soon, an unexpected situation emerged.

AI is beginning to penetrate into various positions within the company, and more and more non-technical employees are starting to develop their own intelligent agents: R&D engineers are incorporating years of accumulated design experience into intelligent agents, maintenance workers are turning common fault handling methods into knowledge bases, and even administrative personnel are using AI to optimize processes

AI is moving away from the IT department and becoming a tool that every employee can use. Mr. Yang sighed and said, "In the past, we thought digitization was just IT, but now we realize that those who understand the scenarios are the most suitable for using AI. ”At present, the company has set the goal of AI transformation to improve overall operational efficiency by 30%, and various landing effects are gradually emerging.

Subsequently, Mr. Yang raised several surprising questions for the author: Since AI can replace so much repetitive labor, can most companies in the future reduce staff? Can all positions from management to grassroots be streamlined and optimized? Can some departments be merged to promote flatter management? With the arrival of the AI era, will manufacturing factories in the future face significant layoffs?

In response to these questions, we have engaged in in-depth discussions and reached a core consensus: if AI and digitization are only used as "scissors" to reduce manpower and cost, enterprises will miss out on the greater value of digital transformation.

Three stages of enterprise digitization

In my opinion, the digitization of enterprises is a continuous process of upgrading dimensions, which can be roughly divided into three stages.

In the first stage, AI employees develop intelligent agents. This is what this small household appliance is doing: developing intelligent agents for people who understand the scene. In the traditional mode, the digital needs of business departments are all piled up in the IT department, resulting in long development schedules and slow response times, becoming the core bottleneck for improving business efficiency. After the popularization of AI low code and zero code tools, the threshold for digitalization of positions has been greatly reduced, and every position can turn its professional knowledge into an intelligent agent. This is the "ignition" stage of digitization: digitization no longer belongs solely to the IT department, but has become a new capability for the company.

The second stage is data connectivity and digital systems. The main job of the IT department in the past was to develop various application scenarios, but when various positions began to use AI and generate data, the role of the IT department underwent a fundamental change - transforming into the system architecture department of the enterprise, stringing together the scattered intelligent agent data, allowing the data to flow and process into a management decision-making system.

This is like building a city, the first stage is for each household to build their own house (develop intelligent agents), and the second stage is to build roads, water, electricity, and pipeline networks (data connectivity). No matter how well the house is built, if the road is blocked, it will always be an isolated island.

The third stage is strategic change. When data truly runs within the enterprise, the strategic perspective of the enterprise will undergo a fundamental change.

In the small home appliance industry, some orders start with only 50000 pieces, which large enterprises do not value - the quantity is too small, and organizing production is not cost-effective. But the market changes quickly, and 50000 pieces may quickly turn into 500000 pieces: if you don't accept them, you miss the opportunity; Got it, the cost can't bear it. Sales oriented companies also dare not place large orders immediately. They need "quick return for small orders", but they are unwilling to give orders to small businesses, fearing quality issues.

If an enterprise relies on an internal unified data system and initially has the ability to split orders and allocate production capacity, it can break down and allocate scattered small orders to different production lines, balance production costs and market demand, and lay the foundation for the subsequent external integration of industrial resources. At this point, the strategy of the enterprise is no longer "how to lower costs a bit", but "how to use data capabilities to reconstruct competitiveness".

Building industry big data capabilities

Completing the internal digital transformation of an enterprise can only solve its own operational efficiency issues. To establish long-term, non replicable competitive barriers, companies must break out of factory boundaries and establish an external big data perspective that covers the entire industry chain. This is also a common transformation blind spot among many Chinese manufacturing enterprises.

By comparing the accumulation of industrial data between China and foreign countries, the gap can be intuitively seen: Bosch in Germany has created a "Failure Mode and Effects Analysis" (FMEA) file for a product, which can reach 12000 pages and has accumulated data for one or two hundred years. From material parameters, process flow to defect modes, every link is supported by massive data. The relevant documents of similar manufacturing enterprises in China are generally only about 80 to 120 pages, indicating a huge gap in industrial digital infrastructure.

The problem facing Chinese manufacturing today is no longer whether it is possible to learn by dismantling other people's products. The things that can be dismantled are basically dismantled, and many fields have even become globally leading. The next step requires' positive development ': using data-driven research and development, rather than imitation driven production. Digitization is precisely the foundation for establishing this positive development capability. Without the accumulation of industrial data, there can be no true positive research and development, which is a lesson that Chinese manufacturing must catch up on.

Tesla entered China and achieved mass production and rapid localization of its supply chain in just one year. Why? Because it has data on the Chinese automotive industry: it knows which parts can be found where, which supplier parameters match, and which link costs can be reduced.

This ability is essentially a 'god's perspective' with big data: companies not only understand themselves, but also understand the entire industry. It can find its position in the coordinate system, benchmark innovation directions within the same industry, and even learn from cross industry technological paths. This ability to "benchmark innovation" is the true competitiveness brought by digitization.

If companies can systematically establish market data, technological routes, and supply chain maps for different categories such as small refrigerators, small rice cookers, small coffee machines, and small air fryers, and then benchmark them with the latest innovations in the industry, they can compete on price and become "industry" at a higher level.

From manufacturing enterprises to platform based enterprises

When small and medium-sized enterprises complete their internal digital transformation and establish a complete industry database, their corporate identity will undergo a fundamental transformation: gradually completing the triple transformation from digital factories, service-oriented enterprises to industrial vertical interconnection platforms, completely breaking away from the positioning of a single product manufacturer.

Digitization is the prerequisite for all upgrades. Through internal three-stage transformation, the enterprise has established its own product database, standardized digital management system, and transformed market capacity, consumer preferences, and component matching system into reusable data assets. Data has become the new core product of the enterprise. Without complete data accumulation, it is impossible to discuss subsequent industrial integration and external services.

By relying on its own data and mature system, enterprises can output services. The above-mentioned rice cooker companies have started to implement related attempts: distributing small orders to surrounding small businesses and sharing their own supply chains - small businesses have flexibility, while large enterprises have supply chain and quality control advantages, creating space for cooperation. Mr. Yang said that once a company has established a complete digital system, it can use order splitting software to accurately distribute small and medium-sized orders to matching small businesses, and be responsible for quality control, standards, and data collaboration.

At this point, large enterprises no longer rely on product manufacturing price differentials for profit, but instead export a complete set of industrial supporting services, transforming their quality control system, research and development capabilities, and supply chain management capabilities into replicable services to earn premiums and transform into industrial service providers.

When enterprises have both massive data assets and standardized service output capabilities, the natural result of further development will be to transform from a single production node in the industry chain to an industrial router that connects the entire industry - a vertical interconnection platform.

The division of labor in today's industry is as follows: thousands of factories are fighting on their own, taking orders based on relationships, quality control based on experience, idle during the off-season, and unable to deliver goods during the peak season. Sales companies dare not place large orders, small factories cannot receive good orders, and the mismatch of resources in the entire industry is very serious.

After the emergence of vertical interconnection platforms, the information flow, order flow, and quality flow of the industry run in the same system. Large enterprises accept, dismantle, and dispatch orders, while dozens or even hundreds of small and medium-sized factories connect to the platform and collaborate on production according to unified standards. Small businesses do not need to build their own quality control systems, maintain research and development teams, or find orders on their own. By accessing the platform, they can acquire these abilities, transforming from a "small workshop that has to do everything on its own" to a "boutique workshop that focuses on manufacturing".

This is not about big fish eating small fish, but about reorganizing the ecology of the entire industry. Xiyin is a typical reference: its essence is not a clothing brand, but a digital supply chain platform. Thousands of small and medium-sized clothing factories have connected to Xiyin's data system, gaining the ability to digitize orders and produce flexibly. There is a symbiotic relationship between Xiyin and these factories.

In fact, many manufacturing companies may take this path. In the future, the core capability of Yang Corporation will be to "organize the production of small household appliances industry". It will become the infrastructure of this industry, like the power grid to appliances and operating systems to software. Although it is not a producer, without it, the efficiency of the entire industry cannot be improved.

This is the true strategic upgrade brought by digitization to the manufacturing industry: it fundamentally changes the business model of enterprises, from earning manufacturing money to earning data money, service money, and platform money, and achieves a qualitative change in the way industries are organized.

When Mr. Yang stepped out of his consideration of whether to lay off employees and began to think about how to use data to connect surrounding factories, the digital awakening of this small home appliance enterprise had just begun.

The road from AI employees to data connectivity, from strategic upgrading to industrial routers, is long, but the change in direction is more important than speed. This is the process of Chinese manufacturing moving from "big" to "strong", and every enterprise needs to undergo a new transformation.

I have always believed that digitization is a new industrial revolution. It not only brings technological iteration and improved production efficiency, but also reshapes the production relations of the entire industry chain. AI can replace repetitive labor, but it cannot replace a company's understanding of the industry; Data can compress costs, but it cannot generate new business models. Treating digitalization as an interface connecting the external world will make enterprises softer and stronger, and their industrial status indispensable.

(The author is the president of the Guangdong Small and Medium sized Enterprise Development Promotion Association and the author of "The Right Path")

(Author Xie Hong)