On September 12, during the "Insight 2035: Industrial Breakthrough and Intelligent Evolution Driven by AI" session at the 2025 Inclusion · Bund Conference, the Antai School of Economics and Management, Shanghai Jiao Tong University, the Bank of China Technology Finance College, China Pacific Insurance Group, L'Oréal China, Lekai Sports, Industrial Bank, and Ant Group jointly launched the industry's first enterprise AI Adoption Maturity Model (AI Adoption Maturity Model, AIM²).
This model introduces five progressive levels, from "point trials" to "AI-native" (L1-L5) for the first time, and provides a systematic framework and implementation path for enterprises to assess and implement AI based on six key dimensions: strategy, organization, data, technology, application, and business. It helps enterprises accurately position their current status, clarify transformation directions, and achieve a value leap from "+AI" to "AI+".
Figure caption: The release of the industry's first enterprise AI Adoption Maturity Model (AIM²)
As large model infrastructure gradually becomes similar, the focus of competition among enterprises is shifting from "racing for computing power and parameters" to "racing for data and scenarios." Whether enterprises can deeply integrate internal data and industry knowledge and apply them to practical business scenarios has become a key differentiator in AI capabilities.
"The first half of AI was about the number of model parameters, and the second half is about the quality of data," said Professor Liu Shaoxuan, head of the AIM² research team, associate dean of the Antai School of Economics and Management, and executive director of the Bank of China Technology Finance College at Shanghai Jiao Tong University.
Currently, many enterprises have gradually moved away from "model anxiety" and shifted toward practical implementation. A few leading enterprises have already deeply integrated AI into their businesses. However, most enterprises still face challenges where AI implementation does not meet expectations. The fundamental reason lies in the lack of high-value, high-quality data.
To address the pain points of enterprises in AI implementation—“not seeing the current situation clearly and not finding the right path”—AIM² constructs a systematic enterprise AI maturity assessment system based on six key dimensions: strategy, organization, data, technology, application, and business.
The research team conducted in-depth surveys on four industries: finance, automotive, health, and retail. They found that each industry has distinct characteristics, and enterprises within different industries are at different stages of AI maturity:
l Financial: Strong data foundation but commercial aspects need further improvement. Enterprises are moving from "decision support" to "autonomous financial intelligence agents." China Pacific Insurance Group built a "digital workforce," upgrading AI from an auxiliary tool to an intelligent agent capable of independently completing tasks, achieving a service leap from "insurance claims" to "risk reduction." Industrial Bank focuses on scenarios that can quickly bring business value, and through building hundreds of intelligent agents, it enables AI to transition from "being able to speak" to "being able to act."
l Automotive: Strong strategic and organizational capabilities but weak data foundations. Enterprises are transitioning from "product intelligence first" to "dual drive of product intelligence and enterprise intelligence." Zero Run Auto maintains an ROI-oriented approach, avoiding excessive pursuit of large models, and achieves full-domain self-research to build a soft-hardware integration solution, creating precise differentiated experiences.
l Health: Most balanced development, transitioning from "mass customization" to "full-cycle proactive health service innovation." Lekai Sports uses AI to operate "smart stores," scheduling resources in real-time and intelligently matching coaches; the Digital Healthcare Division of Ant Group connects multiple resources in the medical ecosystem through technology, significantly improving the efficiency, accuracy, and accessibility of services.
l Retail: Overall in a catching-up stage, needing to upgrade from "workflow improvement" to "consumer-centered experience leap." L'Oréal China places great emphasis on "consumer value creation," actively building an open beauty tech ecosystem. Through external AI skin analysis tools and content automation technology, it rapidly enhances the scale of product experiences.
Figure caption: AIM² divides enterprise AI maturity into five levels, forming an executable AI advancement roadmap
Professor Liu Shaoxuan vividly described: "The six key dimensions together form the 'launch vehicle' for enterprise AI application—the strategy is the command module, determining direction; the organization is the propulsion system, providing support; data is the fuel; technology is the flight control system; application is the navigation route; and commercial value is the target planet." He emphasized that these six dimensions are interlinked, and any weakness would restrict overall development.
Industry professionals pointed out that the core value of AIM² lies in breaking down the complex AI implementation process into orderly stages, providing enterprises with clear evolutionary paths and best practices, avoiding blind exploration. By systematically assessing the six key dimensions, enterprises can accurately position their stage, identify weaknesses, plan paths, and quantify the real returns of AI investments. Ultimately, AI will be advanced from "concept validation" to "scale benefits," building long-term competitive advantages.
In the future, the competition in enterprise AI applications will be a test of systematic capabilities. Only by deeply integrating AI into core value chains can enterprises build sustainable advantages in this transformation, crossing from "technology adopters" to "core drivers."
