Feasibility and Acceptance of Conversational AI for Scheduling Outbound Calls in Home Aged Care: Mixed-Method Service Evaluation (Preprint)

Abstract

<sec> <title>BACKGROUND</title> As populations age and demand for in-home age care increases, home care providers face increasing administrative pressures. One area where productivity gains are likely to be significant via optimisation is rostering and scheduling and appointment management. Conversational artificial intelligence (AI) offers a promising solution to automate structured, non-clinical interactions, potentially reducing workforce burden and enhancing client experience. Despite growing popularity, evidence on feasibility and acceptance among older adults remains limited, and assumptions persist that this population resists digital innovation, particularly AI. </sec> <sec> <title>OBJECTIVE</title> To evaluate the feasibility, client acceptance, and operational implications of implementing conversational AI for outbound scheduling calls in aged home care service delivery. </sec> <sec> <title>METHODS</title> An Australian home aged care provider, Silverchain, conducted a mixed-methods pilot in Western Australia’s Greater Southern region (August–November 2025). AI-driven calls were implemented for appointment confirmations, time changes, and cancellations, with a focus on user acceptance and operational insights rather than immediate efficiency gains. Quantitative data included client eligibility records, detailed call logs (attempts, outcomes, transfers), and complaint reports. Qualitative data were derived from nine semi-structured stakeholder interviews (1 consumer, 1 vendor, and 7 staff). Interviews explored perceptions of usability, workload impact, and future integration. Transcripts were coded thematically. </sec> <sec> <title>RESULTS</title> Of all eligible clients, 86.8% remained in the AI-calls pilot while 13.2% opted out. Across 915 attempted calls with client engagement, messages were successfully delivered in 679 (74.2%). Identity verification failed in 130 calls (14.2%), 39 calls (4.3%) were abandoned mid-call, and only 21 (2.3%) required transfer to the call centre. Complaint rates were negligible (&lt;0.5%). Contrary to prevailing assumptions, older adults demonstrated high receptivity to AI-mediated communication. Thematic analysis revealed three dominant themes: (1) alignment with broader digital transformation goals; (2) perceived potential for future efficiency gains; and (3) recommendations for future improvements to fully realise AI benefits. As expected, no short‑term efficiencies were realised and staff workloads temporarily increased; however, interview participants viewed conversational AI as a viable pathway to future operational improvements, contingent on full integration with core systems. The pilot coincided with low appointment cancellation volumes, constraining full scalability assessment. </sec> <sec> <title>CONCLUSIONS</title> Conversational AI is feasible for managing outbound scheduling calls in home aged care, with high client acceptance challenging myths of digital resistance among older adults. The pilot yielded critical organisational learnings: successful adoption requires robust planning, technical readiness, and alignment with broader digital transformation strategies. These findings can inform future models of care and underscore the potential of AI to support automated calls and sustainable service delivery in aging populations. </sec> <sec> <title>CLINICALTRIAL</title> Not applicable </sec>

Ähnliche Arbeiten