Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Socially assistive robots offer promising opportunities to support older adults in healthcare settings by enhancing communication, reducing loneliness, and promoting emotional well-being. Beyond direct patient interaction, socially assistive robots may also assist healthcare professionals by facilitating information delivery, relieving workload, and improving the organization of care activities. However, ensuring that these systems are both acceptable and usable by older adults and healthcare staff remains a critical challenge, particularly in dynamic hospital environments. Objective: This study aims to evaluate the acceptability and usability of a social robot in a geriatric day care hospital among patients and their informal caregivers and to identify facilitators and barriers to the adoption of social robots in healthcare settings. Methods: Tests were conducted between May 2023 and July 2024, involving 97 participants, including patients (n=65) and informal caregivers (n=32), across three experiments, called waves. The ARI robot, developed by PAL Robotics, was used in this study. Participants were invited to interact spontaneously with the robot as long as they wished, after which they completed usability and acceptability standardized questionnaires. They were administered orally and complemented by open-comments. Results: Quantitative analysis showed a significant increase in usability and acceptability scores across the three experimental waves. and qualitative analysis revealed similar improvements and participants reported greater comfort and satisfaction, particularly after the integration of a large language model. Conclusions: Successive upgrades of the social assistive robot, driven by large-language-model integration, led to significant increases in acceptability and usability scores across the three experimental waves, both in patients and caregivers. Introducing the LLM enabled the robot to engage in dialogues that were more natural, coherent, and context-sensitive. These findings show the value of an iterative, user-centred refinement process when integrating social robots in geriatric care. Clinical Trial: The study was approved by the French national ethics committee: “Comité de Protection des Personnes, CPP Ouest II, Maison de la Recherche Clinique-CHU Angers” (IRB: 2021/20) and was compliant with the General Data Protection Regulation (GDPR) (DPO: 20210114153645, AP-HP register).