Coordination Mechanisms in AI Development: Practitioner Experiences on Integrating UX Activities

Research Background and Issues

• What problems or challenges did the authors identify?
  With the increasing application of artificial intelligence (AI) technologies in products and services, critical challenges in collaboration have emerged, including interdisciplinary cooperation issues between user experience (UX) design and AI development teams. These challenges involve aspects such as the impact of coordination mechanisms on team collaboration and product success, specifically:

  1. The "capability uncertainty" and "output complexity" of AI systems.
  2. Differences in coordination mechanisms between UX and AI teams during AI development, such as the use of output standardization and skill standardization, which may lead to silos, power asymmetries, and delays in the design phase.
  3. A lack of tools and interdisciplinary knowledge to support collaboration between UX design and AI development.

• Why is this issue important?
  As the complexity and impact of AI technologies grow, addressing interdisciplinary collaboration challenges is critical. Effective coordination mechanisms can enhance development efficiency and ensure that the final product meets user needs. Failure to overcome these collaboration challenges may result in products that fail to deliver sufficient user value and ultimately fail in the market.

• Research Motivation and Related Work
  Existing research in this field has proposed using human-centered AI (HCAI) design guidelines and functional prototyping tools to support interdisciplinary collaboration. However, the authors point out that these methods have limited effectiveness at the organizational level in practice. Analyzing this issue from the perspective of coordination mechanisms may provide more systematic and generalizable solutions.

Solutions

• What methods or solutions did the authors propose?
  The authors conducted an in-depth study of collaboration between UX design and AI development teams based on Mintzberg's organizational coordination mechanisms theory (including mutual adjustment, standardization of skills and outputs, and direct supervision). They explored how optimizing team coordination mechanisms can facilitate product design and development. The study provided the following recommendations:

  1. Avoid relying solely on team separation coordination mechanisms achieved through "output standardization" or "skill standardization."
  2. Adopt "mutual adjustment" as the primary coordination mechanism, such as promoting continuous communication and flexible collaboration between teams.
  3. Enhance UX designers' conceptual understanding of AI technologies, for example, by supporting interdisciplinary knowledge exchange through workshops and co-creation design processes.

• What is innovative about this solution?
  The core innovation lies in providing a framework-based perspective grounded in Mintzberg's theory to systematically analyze and optimize interdisciplinary collaboration issues. It identifies coordination mechanisms as the key determinants of these challenges, an area previously underexplored in UX and AI collaboration research. Additionally, the study highlights the risks of standardization-based coordination mechanisms and evaluates the advantages of direct supervision and mutual adjustment mechanisms, deepening the theoretical understanding of AI design-development collaboration challenges.

• What are the implementation steps and key technologies used?

  1. Conducted in-depth interviews with 15 industry practitioners from different companies and roles (including UX designers and AI developers) to collect field data.
  2. Built an analytical framework using Mintzberg's theory and organized and summarized the impact of coordination mechanisms through thematic analysis.
  3. Compared three types of coordination mechanisms in team organization (direct supervision, standardization of outputs and skills, and mutual adjustment), highlighting the different impacts of each mechanism on the design process, power distribution, and resource requirements.

Research Findings

• What specific findings were achieved?

  1. Identified risk markers of standardization-based coordination mechanisms, including team silos, power asymmetries, insufficient user value definition, and a lack of effective user needs exploration in the early development stages.
  2. Demonstrated that mutual adjustment mechanisms lead to more successful team collaboration, including establishing more balanced team power dynamics and improving the interdisciplinary influence of UX design.
  3. Highlighted current resource shortages in AI design, such as the lack of specific AI prototyping tools and effective HCAI design guidelines.

• What are the advantages compared to existing solutions?

  1. Provides a more systematic analysis and resolution of UX and AI collaboration issues, offering an organizational theoretical framework applicable to various team settings.
  2. Proposes practical recommendations based on real-world practitioner experiences and challenges, rather than purely theoretical guidelines or tool development.

• What were the experimental or evaluation results?
  Through the interview study, the applicability and strong explanatory power of Mintzberg's coordination mechanisms theory in addressing team collaboration challenges were confirmed. The research found that mutual adjustment mechanisms outperform standardization-based coordination mechanisms and, in certain cases, effectively address design obstacles posed by AI system capability uncertainty and output complexity.

• Limitations and Future Directions

  1. The study's sample size was limited to 15 interviewees, most of whom were from European companies. Future research should include a broader range of geographic and domain contexts.
  2. The role of management in supporting or hindering coordination mechanisms has not been thoroughly explored and could be a focus of future research.
  3. There is a need to develop AI behavior prototyping tools and investigate how to integrate these tools into the practices of separated teams to optimize standardization mechanisms.

Conclusion

By integrating Mintzberg's coordination mechanisms theory with UX design and AI development practices, this study provides an in-depth analysis of interdisciplinary team collaboration challenges and solutions. The research highlights the advantages of "mutual adjustment" mechanisms in AI design while warning against the high risks associated with team separation achieved through "output standardization." Future research should focus on further promoting resource development, optimizing educational tools, and refining user value definition practices to support the creation of successful AI development processes and product designs.