Year: 2026
Type: article
Study design: Systematic Review
Abstract: The strategic shift toward community-based healthcare, away from hospital settings, is a policy priority internationally. Conversational artificial intelligence (CAI) may support this transition, yet its practical applications remain underexplored. This scoping review maps CAI capabilities and their configuration within community care. We searched PubMed, the ACM Digital Library, and selected grey-literature sources for studies published from January 2023 to June 2025. Of 551 records identified, 29 studies were included. We identified 20 capabilities across five functional domains: Understand (n = 5), Transform (n = 3), Remember (n = 3), Generate (n = 5), and Act (n = 4). Findings suggest that CAI extends rather than replaces human care, supporting access, ongoing engagement, and care coordination. Multi-party and expert-in-the-loop designs emerged as recurring implementation strategies. They also surface design tensions, including tradeoffs between reliability and conversational flexibility, and between the benefits of long-term memory and associated privacy risks. The resulting capability-centered taxonomy offers accessible framework for integrating CAI into community care settings.
Institution: Imperial College London
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W7161465457