Co-designing conversational agents with older people: A scoping review of methods, challenges, and a way forward

Conversational agents are increasingly recognized as potential solutions to the challenges of aging. However, adapting these technologies to the specific needs of the older adults requires effective design methods that integrate end-users and relevant stakeholders throughout the development process. Co-design methods have gained popularity in this context, but the criteria for evaluating their effectiveness remain limited to basic measures of acceptability and ease of use, with little focus on their actual impact on design outcomes. This study addresses this gap by aiming to: 1) identify the structures of co-design methods tailored for older adults, 2) analyze their strengths and weaknesses, and 3) provide recommendations to optimize future applications.

To achieve these goals, we conducted a scoping study using the PRISMA methodology. Searches were carried out on PsycInfo, Google Scholar, PubMed, ACM, IEEE, Web of Science and DBLP. Of the 982 articles identified, 27 projects were selected. The analysis revealed great variability in methods and highlighted the lack of consensus on optimal co-design structures for this population. In addition, most studies lacked measures to assess the influence of method diversity on design outcomes and participant experience.

These findings underscore the need for a scientifically grounded design framework and guidelines to support best practices in co-design for aging-related conversational agents. Establishing such standards would reinforce methodological rigor, enable better comparability across studies, and allow for more accurate measurement of co-design impact on product effectiveness and user satisfaction.