Fosun Group AIGC Campaign

1,590,000

In terms of strategy and mechanism design, I led the construction of Fosun’s “3+2+2” Generative AI Capability Matrix and launched a unified AI platform to support early-stage applications across subsidiaries. The platform’s token usage grew from 3.83 million (May 2023) to 304 million (November 2023). I designed and implemented the Group’s AI Value Assessment System, evaluating applications along four dimensions—business, user, technology, and long-term benefits—and introduced a reward-and-penalty mechanism to incentivize innovation. I also chaired the establishment of Fosun’s AI Subcommittee, comprising 12 internal members (CTOs of key subsidiaries) and 9 external members (professors from Tsinghua University, the University of Science and Technology of China, and the Chinese Academy of Sciences, along with experts from Alibaba, Microsoft, NVIDIA, and Ping An). To foster exchange, I organized 12 high-level AI seminars and conferences, including closed-door workshops on AI in pharmaceuticals, insurance, and cultural tourism, as well as the Fosun Global AI Ecosystem Conference. Meanwhile, Fosun joined the Shanghai Artificial Intelligence Association and partnered with the China Academy of Information and Communications Technology and the Zhongguancun AI Research Institute.

The outcomes of this campaign were substantial. In 2023, 17 subsidiaries launched large model application projects, creating RMB 83.48 million in business value. In 2024, the scope expanded to 27 subsidiaries, delivering RMB 123 million in cost savings and revenue gains. In total, 126 new AI Agents were deployed across R&D, product development, marketing, and service operations.
• R&D: Fosun Pharma’s PharmAID decision-support system improved strategic decision efficiency by 50% and is expected to increase drug development success rates by 5%. Fosun Henlius, through its AIDD platform, developed small molecules targeting c-myc RNA, potentially shortening R&D timelines by 70% and reducing compound testing and trial costs by one order of magnitude.
• Product development: Artery Network’s multimodal medical large model reached 200 million annual diagnoses and was recognized in the 2024 Forbes China Top 50 AI Companies. Shanghai Ganglian, in partnership with the Shanghai AI Laboratory, released the Zongshi Commodity Large Model and Xiaogang Robot 2.0, reducing operating costs by RMB 9 million and generating RMB 3 million in additional revenue. Babytree launched the Mika Mother-Infant Large Model and Hourlight 2.0, increasing revenue by RMB 4 million.
• Marketing: Club Med deployed the TED Copilot system across 25 countries and regions, achieving 80% customer satisfaction and reducing sales staff time by 50%. Yuyuan Jewelry implemented multimodal product search across 589 stores, increasing conversion rates by 1%. Its 13-billion-parameter Jewelry-LVLM model optimized gold jewelry classification and improved supply chain efficiency by 30%. At Yuyuan’s Bund Financial Center (BFC), AIGC-based art installations attracted 15 million offline visitors. In Portugal, Fidelidade Insurance’s GAMA-AI solution improved insurance sales conversion rates by 10%.
• Services: Foshan Chancheng Hospital used AI-driven outbound calls (20,000 in total) to bring back 2,300 patients for consultations and successfully intervened with 78 potential critical cases. Jianjia Medical’s AI quality-control system reduced case auditing from minutes to real-time full coverage. Debon Securities’ AI Investment Radar increased client retention by 50% and delivered excess returns to 90% of clients.

The campaign not only advanced Fosun’s organizational transformation toward a “Human + AI” model, but also built a sustainable AI innovation ecosystem across its businesses. Its achievements were showcased as best practices for enterprise AI transformation at the World Artificial Intelligence Conference (WAIC) and the Silicon Valley Generative AI Summit. The project received international recognition, winning the 2024 Microsoft Global Award for Pioneering Applications of AI.