A Multimodal Conversational Chatbot for Emotional Support and Daily Assistance in Older Adults: Design, Development and Pilot Usability Evaluation

Background/Objectives: Population ageing and the increasing prevalence of loneliness and social isolation represent major public health challenges. Digital health technologies, including conversational agents, have emerged as potential tools to support emotional well-being and daily functioning in older adults. This study presents the design and exploratory pilot evaluation of UjaBienestar,a multimodal conversational chatbot aimed at providing accessible emotional and practical support. Methods: A web-based multimodal system integrating text and voice interaction was developed using a Django backend and a Rasa-based conversational engine. The system was designed following user-centred and accessibility-oriented principles for older adults. A pilot usability and feasibility study was conducted with 10 participants using task-based interaction, observational data and pre/post-questionnaires. The exploratory evaluation focused on preliminary user perception, accessibility, and interaction feasibility rather than on statistically generalizable or clinically validated outcomes. Results: Participants reported high levels of perceived usability, accessibility, and acceptability. Multimodal interaction, particularly voice support, was positively valued. Users reported subjective perceptions of companionship during interaction within this pilot context. Initial barriers were mainly related to onboarding and first-time use. Conclusions: The findings suggest that the proposed system is a feasible and acceptable digital health support tool for older adults. While the results are preliminary, they highlight the potential of multimodal conversational technologies for supporting perceived emotional well-being, accessibility, and daily assistance in ageing populations. These findings are based on perceived user responses and do not represent clinically validated outcomes. Further large-scale and longitudinal studies are required to assess clinical and psychosocial impact.