A paper validating the effectiveness of CocktailAI, a generative AI-powered medical documentation solution, has been published in PLOS Digital Health. A joint study with Kyoto University Hospital conducted an empirical evaluation of a clinician-in-the-loop workflow, indicating the potential for improvements in work time and document quality.
A joint study by Kyoto University Hospital and フィッティングクラウド conducted an empirical trial to validate the utility of CocktailAI, and the results have been published as a paper. The paper was published in the international academic journal PLOS Digital Health on 27 October 2026.
The study compared 3 clinical validation groups: the Clinician-in-the-loop group, in which physicians revised AI-generated drafts; the Clinician-only group, in which physicians created documents manually; and the LLM-only group, using AI alone. The report indicated a trend toward higher document quality when the workflow included physician review.
The primary analysis suggested a 59.4-second reduction in the time required to prepare discharge summaries. It also noted that AI-only output failed to meet the quality standards. The study is said to indicate that physician intervention may enable both improved document quality and reduced documentation burden.
Going forward, the company aims to expand integration with in-hospital documentation systems and establish flexible implementation environments to promote broader use in healthcare settings.