德扑之星

The AI for Process: Enterprise-Level Intelligent Process Transformation Blue Paper, jointly developed by Digital China, Deloitte China, and the China Academy of Information and Communications Technology (CAICT), focuses on the application and practice of enterprise-level AI in process digital-intelligent transformation. It establishes a full-cycle guidance framework from strategic planning to technological implementation, providing enterprises with a systematic solution to overcome the challenges of perception, methodology, and practice in AI adoption.

Over the next three to five years, the most significant shift in organizational structure will be a stronger move toward flattening.This transformation will directly reshape corporate decision-making: instead of relying on layer-by-layer outputs from different departments, management decisions will increasingly be based on the direct interpretation of data—and in many cases, on AI-driven analysis. As a result, decision chains will be dramatically shortened, and organizational structures will become significantly more streamlined.

—————— Lu Wenyan, Vice President, Shanghai Oriental Digital Commerce Co., Ltd.

Although there are not yet clear cases of AI models directly generating revenue, their impact in reducing enterprise costs—both directly and indirectly—is already evident. Overall, the return on investment (ROI) can be considered relatively reliable.

—————— Xu Dong, Vice President of Alibaba Cloud & General Manager of Tongyi Large Model Business

Choosing a reliable supplier for long-term collaboration can save significant hidden and communication costs. In particular, replacing suppliers in core business areas often requires a two- to four-year adjustment period, during which hidden communication and adaptation costs are substantial. Therefore, we prefer to establish a long-term, tightly integrated partnership with our collaborators.

—————— Wang Jinnan, General Manager of Digitalization and Information Technology, Swire Properties China

We can confidently predict that tasks such as data handling—transferring data from one system or offline source to another a