Not all enterprises are suitable for implementing AI.
Currently, AI technology is developing at an unprecedented speed, permeating all industries and becoming the "new favorite" that many enterprises are vying to pursue. From intelligent customer service improving customer service efficiency to intelligent production optimizing manufacturing processes, AI seems to paint a development blueprint full of infinite possibilities for enterprises. However, after getting deeply involved in this field for nearly half a year, Lao Yang found that not all enterprises are suitable for embracing AI in this AI technology frenzy. If enterprises blindly follow the trend and introduce AI, some may fall into the dilemma of huge investment but meager returns.
Why is that?
Essentially, an enterprise's introduction of an AI project is also a digital project. Looking back at the past digital transformation construction of enterprises in recent years, it's not difficult to find that even some enterprises with sufficient budgets still have numerous cases of failed digital transformation. Why is this so? Lao Yang summarized the following five core reasons for failure:
1. Strategic defocus: unclear strategy and ambiguous purpose;
2. Organizational confrontation: there has always been an implicit or explicit game between business departments and information departments;
3. Competency gap: employees lack digital capabilities, but many enterprises are not aware of this. Many enterprise leaders still promote it with a blindly optimistic attitude. Relevant research data shows that when the proportion of employees over 50 years old in an enterprise exceeds 40%, the training cost doubles. When the proportion of employees with an educational background below junior college is 30%, the data entry error rate exceeds 25%;
4. Management lag: lack of supporting management mechanisms and processes, unable to adapt to the changes brought about by new technologies; this leads to a sharp increase in later - stage costs. Relevant data shows that when an enterprise lacks data standards, the later - stage governance cost will account for 28% of the entire digital budget;
5. Accumulation of technical debt: blindly pursuing technological advancement while ignoring the adaptability between technology and business, resulting in a bloated and difficult - to - maintain system. Enterprises may have to completely rebuild it, causing the previous investment to go to waste.
Enterprises already face such challenges in digital projects. Moreover, as a more complex and disruptive technology, AI puts forward higher requirements for an enterprise's overall collaborative ability and strategic execution. Therefore, Lao Yang believes that before introducing AI, enterprises must calmly evaluate their own conditions and clarify the boundaries of technology application. Blindly chasing hotspots will only lead enterprises into the misunderstanding of "using AI for the sake of AI". Therefore, Lao Yang suggests self - evaluation from the following four scenarios:
First, Digital foundation:
Whether an enterprise has good data governance capabilities and information infrastructure is the foundation for AI application; business matching degree: whether AI technology is highly compatible with the enterprise's core business scenarios and can truly solve practical problems. For example, whether the data quality meets the standards and whether the system and data silos are connected are the keys to determining whether AI can deliver value;
Second, Business scenarios:
Currently, some traditional enterprises seem to have entered a misunderstanding when introducing AI, that is, introducing AI is only for laying off employees and reducing costs, which is obviously a misinterpretation of the value of AI. The core value of AI lies in "increasing efficiency", rather than simply "reducing the number of employees". In the process of adapting to business scenarios, AI is more applicable to fields that require a large amount of repetitive labor, high - precision decision - making support, or real - time response, such as the quality inspection process in the manufacturing industry, risk assessment in the financial industry, and intelligent recommendation in the retail industry.
Third, Organizational ability:
AI technology puts forward higher requirements for an enterprise's organizational ability. It not only requires a cross - departmental collaborative mechanism but also needs compound - type talents, especially those who can integrate technology, business, and management. Research data shows that for an enterprise to truly carry out AI - related work, the proportion of compound - type talents needs to reach more than 15%; otherwise, the project success rate will be significantly reduced. The most crucial factor is the maturity of change management. When facing the process re - engineering and organizational changes brought about by AI, enterprises often lack a systematic coping strategy. Many enterprises ignore employees' adaptation and participation when introducing technology, resulting in difficulties in promoting changes and even internal resistance.
Fourth, Economic viability:
The purpose of enterprises introducing AI is to reduce costs and increase efficiency. However, if they only focus on short - term cost savings, they may easily ignore the long - term value brought about by AI applications. In fact, the real economic value of AI is reflected in efficiency improvement, quality optimization, and business model innovation. Therefore, when evaluating the input - output effectiveness of AI, enterprises should pay more attention to long - term benefits rather than short - term savings. The ROI cycle should be planned based on a three - year benchmark. At the same time, at least 30% of flexible space should be reserved in the budget to cope with the uncertainties brought about by technological iteration and application deepening.
The above four dimensions are important factors that enterprises cannot ignore when introducing AI. In the process of promoting the implementation of AI, enterprises must conduct systematic planning in combination with their own development stages and resource capabilities. If these key factors are ignored, the introduction of AI will not only fail to create the expected value but may also increase the burden on enterprises and even trigger a series of problems.
Currently, most traditional enterprise bosses are in a state of AI anxiety. On the one hand, they are eager to enhance the competitiveness of their enterprises by introducing AI. On the other hand, they are full of doubts about the actual value, application path, and input - output of AI. Now, let's talk about the key point of this article: Which enterprises are not suitable for introducing AI technology? Lao Yang summarized as follows:
First, Those with a weak digital foundation:
Such enterprises often lack a complete information system, have serious data silos, and have a low degree of business process standardization. They even have difficulty in accumulating basic data to support the training and application of AI models. Research data shows that for enterprises with an electronic rate of key business data less than 50% and more than three data silos, the success rate of AI application is less than 20%.
Second, Those dominated by non - standard business
For example, high - end custom clothing enterprises design and produce each piece of clothing according to the unique needs of customers. From fabric selection, style design to size cutting, there are extremely high customization requirements. Data shows that in such enterprises, the proportion of standardized data is less than 30%. AI models cannot obtain enough effective input and cannot form reusable intelligent capabilities. If such enterprises blindly introduce AI, they will not only fail to achieve the expected results but may also cause greater cost waste due to model misjudgment and resource misallocation.
Third, Those with stubborn management inertia
Such enterprises often have a serious path dependence. Management is accustomed to the traditional management model and has a natural resistance to change. Even if they introduce AI technology, it will be difficult to implement AI applications due to the mismatch between decision - making mechanisms, organizational culture, assessment systems, and intelligent requirements. Research data shows that when the average age of middle - level managers in an enterprise exceeds 45 years old, the technology acceptance rate decreases by 37%, and the resistance to change promotion increases by more than 50%. In this case, AI will not only fail to achieve the expected effectiveness but may also intensify internal contradictions and hinder the normal operation of the enterprise.
Fourth, Those sensitive to costs for survival
Currently, most enterprises are in a survival mode. Under the pressure of a tight capital chain and high costs, these enterprises often cannot afford the high initial investment of AI technology, including hardware procurement, software development, and talent introduction. Data shows that for enterprises with an annual revenue of less than 500 million and an AI budget of less than 1% of the revenue, or a gross profit margin of less than 20% and an ROI requirement of less than 18 months, the failure rate of AI projects is as high as more than 65%. Such enterprises should give priority to ensuring healthy cash flow and strengthening basic management rather than blindly chasing technological hotspots.
Fifth, Those with technological illusions
Such enterprises often have overly high expectations for AI technology, thinking that AI can solve all problems and even expecting to quickly disrupt the industry pattern by introducing AI technology. At the same time, they think that purchasing hardware or a large - scale model means having intelligent capabilities. In fact, such enterprises often lack a clear understanding of the boundaries and implementation paths of AI technology. They neither have a clear plan for application scenarios nor the corresponding technology adaptation ability. Data shows that more than 60% of AI project failures are due to unclear demand definition or the disconnection between technology and business.
So how should enterprises scientifically promote AI transformation?
Lao Yang suggests the following:
1. First digitize, then AI - enable, consolidate the data foundation, and unblock the data chain to ensure that data can be obtained, quantified, and analyzed.
2. Make rational decisions and act within your means, choose the entry point for AI application according to the actual development stage of the enterprise, and avoid blindly following the trend.
3. Start with small scenarios, verify the value and then gradually expand the application scope to avoid difficulties in evaluating the effects due to excessive initial investment.
4. Strengthen organizational collaboration, promote management innovation, cultivate a team of compound - type talents, and improve the enterprise's adaptability to new technologies.
5. Pay attention to technology adaptation, select appropriate AI tools and models according to the enterprise's own business needs, and avoid blindly pursuing technological advancement while ignoring the feasibility of implementation.
6. Continuously iterate and dynamically optimize, summarize experience and adjust strategies continuously in the process of AI application to ensure that technology application and enterprise development are upgraded synchronously. At the same time, establish a scientific evaluation mechanism to form a closed - loop feedback from data quality, model performance to business effectiveness, and improve the sustainability and scalability of AI projects.
7. Strengthen risk awareness, identify and prevent potential technological, legal, and ethical risks brought about by AI applications in advance to ensure that technology use is compliant, transparent, and controllable.
8. Deeply integrate AI applications with corporate culture, improve the awareness and recognition of all employees of intelligent transformation, and create an open, inclusive, innovative, and enterprising organizational atmosphere.
9. Be down - to - earth and gradually promote AI transformation, rather than blindly following the trend.
Finally, Lao Yang wants to say that although AI is powerful, not all enterprises are suitable for embracing it. Technology is just a means, and commercial value is the goal. Rational decision - making can truly enable AI to empower enterprise growth.
This article is from the WeChat official account "Xiangjiang Digital Review" (ID: benpaoshuzi). The author is Lao Yang. It is published by 36Kr with authorization.