Through the AI fog, how can enterprises move from "+AI" to "AI+"?
In the past few years, few words have been as eye - catching as "AI" in the industry. Almost all industries are discussing how to use AI to improve efficiency and quality, and numerous pilot projects have emerged in response. However, behind the boom lies a sobering reality.
On the surface, AI seems to be everywhere, but in fact, the number of applications that are truly implemented and create value is still extremely small. Although everyone is talking about the necessity of AI, many enterprises are repeating the same mistakes in their AI transformation. They only add intelligent tools in some local processes to achieve limited cost reduction and efficiency improvement. In essence, this is simply making a simple addition to the original enterprise construction, which is a typical "+AI" way of thinking.
Data from the "Survey Report on the General Situation of Enterprise AI Applications in 2025" shows that the primary reason for 50% of the surveyed enterprises to introduce AI is "insecurity". They recognize the trend of intelligentization, but few have a clear understanding. The underlying logic remains unchanged, which instead exacerbates this "insecurity".
High - level investment often fails to bring equivalent returns. The actual results of most enterprises' AI application implementation still fall short of expectations. The deeper reason behind this is that enterprises always stop at "+AI".
To solve this pain point, the Antai College of Economics and Management at Shanghai Jiao Tong University and the Bank of China School of Science and Technology Finance, together with a group of industrial partners such as China Pacific Insurance Group, L'Oréal China, Leke Sports, Industrial Bank, and Ant Group, jointly released the industry's first Enterprise AI Adoption Maturity Model (AI Adoption Maturity Model, AIM²). The AIM² joint research team led by Professor Liu Shaoxuan, the Executive Dean of the Bank of China School of Science and Technology Finance at Shanghai Jiao Tong University and the Deputy Dean of the Antai College of Economics and Management, conducted in - depth research on typical enterprises in four industries: finance, automotive, health, and retail, and formed a report.
For enterprises, it provides a brand - new way of thinking and methodology. It truly internalizes AI into various aspects such as enterprise strategy, operation, and technology, and ultimately paves the way for the "AI+" path.
The development of AI has entered the second half. In the fierce - fought battlefield, there are still opportunities to gain large - scale returns. The question is, how can AIM² help enterprises seize this opportunity? How can enterprises truly achieve the leap from "+AI" to "AI+"?
01. Why do enterprises always stop at "+AI"?
Enterprises have flocked to the AI wave, but not all of them can reach the shore.
In fact, only a few enterprises with a high maturity level in AI application can deeply integrate AI into their business. More enterprises are lost in the "fog" of AI transformation. In the report "The GenAI Divide State of AI in Business 2025" published by the research team of the Massachusetts Institute of Technology, it is mentioned that a large number of enterprises invest in GenAI pilot projects but find it difficult to translate them into actual productivity improvement and business transformation. Data shows that 95% of organizations get zero return from GenAI projects, and only 5% of organizations successfully integrate GenAI tools into their workflows on a large scale.
Why is there such a huge contrast? Looking across the entire industry, there are numerous chaotic situations when enterprises apply AI: taking winning the "Hundred - Model War" as the goal, but lacking measurement standards for transformation and improvement; being unable to establish a perfect reuse mechanism, so that success becomes accidental; and also lacking a unified measurement standard, making decisions based on intuition rather than scientifically verified results. The causes at multiple levels have led to a wide - spread systematic failure.
Looking back, what holds enterprises back is not the AI technology itself, but the industry's general stay in the "+AI" perception. There is a lack of a systematic framework, emphasizing technology while neglecting application, making it difficult to deeply embed AI into business. Imagine, if the application of AI cannot form a complete solution and cannot clearly answer what specific and measurable value AI can ultimately bring to the enterprise, then this kind of attempt is at most a routine technology test, which is difficult to help most enterprises cross the gap from pilot projects to large - scale implementation.
An enterprise's real application of AI should be a complete development blueprint that integrates AI with business, including clear strategic planning, effective technology implementation plans, and accurate evaluation of AI's commercial value. At the "Insight 2035: AI - Driven Industrial Breakthrough and Intelligent Evolution" forum of the 2025 Inclusion · Bund Conference, Professor Liu Shaoxuan summarized the core of the problem: applying AI should be a relay race from the "+AI" to the "AI+" business model.
02. In the future, how can enterprises achieve "AI+"?
True "AI+" means fundamentally changing the starting point and truly achieving "AI - native" application.
The native AI application methodology is almost subversive. It requires enterprises not to simply add AI functions, but to take AI as the core driving force, and design the architecture and interaction logic around AI capabilities. This transformation enables AI to officially transform from a tool to an engine, driving enterprise strategy and model innovation.
When AI moves from the periphery to the core, its impact also expands from point to area step by step. From production, supply chain to sales and customer service, AI fundamentally changes the operation logic of enterprises and creates new value points.
Data determines the upper limit of AI application capabilities. As the first insurance company deeply involved in the field of "medical insurance + commercial insurance" data integration and direct - claim settlement, China Pacific Insurance embeds AI applications into the "big health care" and "insurance + service" ecological scenarios closely related to national economy and people's livelihood. It successfully integrates multi - scenario data such as medical, meteorological, and Internet of Things data, and occupies a strategic high - ground in the future competition of the big health care ecosystem. In addition, since 2023, China Pacific Insurance has proposed to build "digital labor in the insurance field", upgrading AI from an "efficiency tool" to a "digital employee" and systematically reshaping the operation model.
When AI truly becomes an inherent ability of an enterprise, its value will further overflow. Enterprises can use "AI+" to guide long - term competition towards ecological construction rather than short - term technology races.
The practice of Ant Digital Medical Health in the past few years has confirmed this: its AI health butler AQ reflects the future trend that spans the cycle. It can jump out of the focus on a single scenario application, deeply connect the multi - party resources of the medical ecosystem through technology, and build a closed - loop ecosystem of medical services. The entire business department is an autonomous intelligent agent, which realizes agile collaboration and efficient decision - making across businesses and systems through a unified "perception, decision - making, action, learning" closed - loop mechanism.
It is worth noting that from the insights presented by AIM², the ultimate direction of AI application is that enterprises gradually evolve into "enterprises as intelligent agents". In the future, AI will become the core engine driving enterprise self - evolution. At that time, an enterprise's competitiveness will no longer depend on scale and resources, but on whether it can start from the AI - native logic, continuously learn, adjust, and evolve, and maintain vitality in the technological wave.
Therefore, the model transformation towards "AI+" brought by AIM² is not only related to an enterprise's current development path, but also to its position in the future industry pattern, and is also the key node for an enterprise to cross the cycle. Then, how to use AIM² to achieve maximum benefits?
03. Find the "compass" on the leapfrog path
How to move from "+AI" to "AI+"?
AIM² innovatively constructs a "Five - Level, Six - Dimension" system, providing a clear evolution path for enterprises to apply AI.
Under the system of five levels (L1 - L5) and six dimensions (strategy, organization, data, technology, application, and business), in addition to having an evaluation value for enterprises, the AIM² model is also a "funnel - type" action guide from strategic screening to value realization, guiding enterprises on how to systematically identify, screen, and implement high - value AI application scenarios. This means that there is a comprehensive and reliable solution from establishing the goal awareness of "AI+" to actual implementation.
The five levels set by AIM² clearly define the maturity ladder of enterprise AI application, reflecting the strategic evolution of enterprises from using AI as an auxiliary tool (the "+AI" model) to reconstructing their business with AI as the core (the "AI+" model).
At the same time, AIM² sets up six inter - related key dimensions from point to area, covering the complete closed - loop from top - level design to bottom - level foundation and then to final value realization, prompting enterprises to focus on platformization, standardization, and reuse rather than repeating the "reinvention of the wheel", thus significantly reducing the marginal cost of enterprises.
In addition, AIM² is designed for the implementation of enterprise AI applications. When the model was initially constructed, the differences between industries were taken into account. Professor Liu Shaoxuan said in an interview with 36Kr that on the one hand, considering the general standards across industries, different indicators are needed for balance; on the other hand, the cycle and process of AI technology diffusion in different industries are also considered. Therefore, multi - dimensional balanced indicators and stage - specific indicators highlighting the cycle are particularly important.
Especially when the market has returned to rationality, enterprises need a comprehensive evaluation when making decisions, systematically avoid risks and pitfalls, and gradually plan the development path from "+AI" to "AI+". Through AIM², we can gain a more detailed insight into the AI application status of enterprises in different industries.
For example, for traditional financial institutions, in the initial stage of AI application, it is not necessary to pursue subversive changes. Instead, they should adopt a practical strategy to enter the core areas and optimize key processes. Shanghai Bank focuses on risk management, a core area of finance, and uses machine learning to accurately evaluate credit risks. Based on high - quality data, it achieves precise optimization of pre - loan approval and post - loan early warning. In addition, ecological cooperation is also an important path for the banking industry to quickly improve its AI capabilities. Shanghai Bank actively cooperates with external technology companies, introduces mature systems such as intelligent investment advisors and intelligent risk control, and quickly verifies the actual value of AI in business scenarios.
In the health industry, the challenges of massive data and high concurrency drive the continuous upgrading of technological infrastructure. Meinian Onehealth systematically uses AI technology to reshape the whole - process service. Based on distributed database technology, it upgrades the new - generation intelligent physical examination management system. It not only breaks through the performance bottleneck but also realizes AI - driven service model innovation, promotes the transformation of personalized health services, and achieves the transformation from one - time physical examination services to full - life - cycle health management.
The transforming automotive industry can rely on accurate strategic choices and in - depth technological exploration to avoid the trap of blindly following the trend. The emerging automotive brand Leapmotor has been strictly following the ROI - oriented AI development path, clearly focusing on two core areas: intelligent driving and intelligent cockpit. It efficiently uses limited computing power to achieve precise business scenario optimization and develops expert models deeply adapted to specific scenarios. In the future, it will achieve intelligent transformation under the "AI pragmatism".
Intelligentization is not simply a technical addition, but a fundamental reconstruction of the organization, business model, and ecological cooperation. In the retail industry, L'Oréal has achieved in - depth adaptation of global AI technology to the Chinese market, built a local data center, and formed the ability to identify trends, virtual makeup try - on and other multi - scenario capabilities to feed back the innovation path of the global headquarters. At the same time, it actively builds an open - style beauty technology ecosystem, jointly promotes the rapid concept verification of generative AI and intelligent interaction technology with start - up companies and universities, and continuously strengthens the brand's competitiveness in technological innovation. This also inspires the industry that embedding in the existing ecosystem is more commercially efficient than building a self - owned platform.
The deep integration of humans and machines can jointly create new businesses and growth models, achieving the effect of 1 + 1>2. And this simpler and more efficient AI application model can enable enterprises to move from shallow - level "tool - efficiency improvement" to deep - level "productivity transformation". The AI transformation of Leke Sports, a leading Internet platform in the fitness industry, is a vivid example. In store management, while using an advanced data collection and analysis platform, Leke Sports introduces an AI inspection system and sets up an AI customer service, greatly improving the store operation efficiency through the integration of humans and machines.
04. Calmly embrace the technological wave
Enterprises always inevitably fall into anxiety when facing new technological waves. This is especially true for AI. In the past few years, this technology has undergone several rounds of iterations. Facing the unpredictable future challenges, what potential does "AI+" still have?
Technology solves current problems, while the ecosystem can cross the cycle. When more and more enterprises have a systematic methodology to implement AI strategies and build a balanced and collaborative six - dimensional ability, the AI application ecosystem in all industries will change. What AIM² points to may not only be the breakthrough of a specific enterprise, but also the overall ecological optimization of the industry.
In other words, the future trend of AI application not only guides enterprises to "do the right thing", but also implicitly plans the "road network" of the industry.
Facing the new technological wave, there are both cautious thinking and bold attempts. Every investment is an investment in the future, which tests both patience and foresight. We believe that those enterprises that continuously engage in AI application practice will ultimately gain "wisdom" beyond intelligence, maintain foresight in the fog, forge resilience in the wave, and create their own sustainable competitiveness.
If you are interested in the AIM² model and the "Report on the Maturity Model of Chinese Enterprise AI Applications", you can search for the "AIM2" mini - program on WeChat to view the full text of the report.