Recent developments of artificial intelligence (AI) have open up significant opportunities for organizations to adopt human–AI collaboration, enhancing the efficiency of complex collaborative decision-making. Within this hybrid human–AI team paradigm, the concept of AI teammates is gaining increasing recognition. Despite their practical potential, organizations still face various challenges when introducing AI teammates, including insufficient consideration of user needs, unclear application scenarios, limited use of human experiential knowledge, and concerns around privacy and ethics. To address these issues, this study adopts the Design Science Research (DSR) methodology and incorporates principles from Human-Centered AI (HCAI) to propose a structured method for analyzing and designing AI teammates in complex collaborative decision-making contexts. The method emphasizes active involvement of business staff and systematically supports the acquisition and structured representation of experiential knowledge, facilitating the effective design and deployment of AI teammates. The feasibility and effectiveness of the method are demonstrated through experimental validation. Theoretically, this research contributes to the methodological foundations of HCAI and advances the integration of human experiential knowledge into AI systems. Practically, it offers a business-staff-driven design approach that enhances organizational intelligence and supports the adoption of AI technologies in complex decision-making situations.