Artificial intelligence and clinical work: Designing GenAI mentoring to support decision making in cognitive rehabilitation
Sharon Zlotnik, Meira Levy, Ethan Hadar, Irit HadarBackground
Occupational therapists working in cognitive rehabilitation may face high cognitive demands, and professional isolation. Although AI is used for information retrieval and decision support, its role in supporting reflective clinical reasoning remains insufficiently conceptualized. As scholarly attention shifts from automation toward augmentation of professional judgment, the next step is to understand how AI may support reflective reasoning in complex rehabilitation work.
Objective
To examine key challenges in cognitive rehabilitation and explore how these needs may inform a conceptual design for GenAI-mediated mentoring within clinical work.
Methods
A qualitative study grounded in Design Thinking methodology was conducted with 30 female occupational therapists (age range 24–64 years, M = 41.26, SD = 10.35) with wide range of clinical experience (less than one year to more than 20 years) from diverse settings. They completed a demographic questionnaire and participated in two Design Thinking workshops (n = 14; n = 16). Seven participants subsequently engaged in follow-up interviews to refine design implications. All sessions were audio-recorded, transcribed, and analyzed thematically.
Results
The thematic analysis yielded four themes: (A) The Therapist's role (B) Challenging therapist-client interactions, (C) Loneliness of the Therapist (D) Expectations for guidance or support aspects. These findings informed the conceptualization of a GenAI mentoring framework grounded in real-world clinical demands.
Conclusions
The findings support the design of generative AI systems as dialogic, context-sensitive tools intended to strengthen professional judgment, warranting further examination of their integration within real-world clinical workflows.