Adapting Team Science Tools for AI Collaboration: Lessons from INVITE

Our proposed session examines how large interdisciplinary organizations, such as NSF’s INVITE AI Institute, use monitoring and evaluation to strengthen collaboration. Over two years, we adapted and implemented the Team Science Survey, derived from TeamSTEPPS, to evaluate team dynamics across scientists, educators, and technologists. Although rooted in healthcare, the tool’s principles proved transferable, generating evidence on communication strengths, coordination challenges, and the conditions that enable effective interdisciplinary work. These findings informed strand‑level improvements, shaped leadership decisions, and demonstrated the value of mixed‑method evaluation in complex research environments. The session will share how evidence from the survey guided action, why multiple evaluation methods are essential for understanding team science, and how systematic monitoring can enhance collaboration in national‑scale research institutes.