AI-Assisted Cognitive Support for Caregivers: A RAG and EUD Framework for Geriatric Care

The increasing prevalence of cognitive decline among the ageing population demands scalable, user-friendly tools that empower healthcare professionals to deliver personalised support. This paper introduces a framework that integrates End-User Development (EUD) principles with Retrieval-Augmented Generation (RAG) techniques to enable geriatric caregivers to design, deploy, and adapt customised cognitive intervention plans. Through a web-based platform, caregivers can profile patients, administer screening questionnaires, and automatically generate tailored cognitive exercises delivered via a mobile application featuring a conversational agent. The agent guides patients through daily cognitive tasks while adapting to individual profiles and progress. A preliminary usability evaluation with healthcare professionals assessed the system's learnability, usefulness, and integration potential. Results suggest that the proposed approach can support caregivers in delivering adaptive cognitive care while lowering their technical and operational burden. This work offers promising insights into using AI-assisted, caregiver-driven tools in geriatric healthcare settings.