Abstract Falls are a significant issue for older adults, causing physical, emotional, and financial strain. Research indicates that regular and early fall screenings can reduce fall incidents and their severity. However, this test usually requires supervision by healthcare professionals, limiting its accessibility, especially in underserved and rural areas. Current technologies are often costly, invasive, and complicated to use independently, with many relying on self-reported data and needing professional oversight for comprehensive fall risk assessment. ForesightCares presents AVA as the first AI-powered, video-based mobile app for reliable at-home fall risk assessment. AVA, a software-only solution, uses built-in cameras and CPU processing capabilities of smartphones for intelligent processing. It is primarily designed for caregivers of older adults, including family members and staff in Independent Living communities, to perform the assessment. AVA differentiates itself by using AI to assess gait, balance, and strength, aligning with CDC guidelines. Its deep learning-based visual mobility assessment technology operates entirely on-device, enhancing privacy and accessibility while avoiding external servers or cloud reliance. This makes AVA more accessible and affordable for a broader range of users. However, the current version of AVA still requires validation and customization for independent use in older adult living environments. This Phase I study aims to validate AVA’s AI and evaluate its feasibility and acceptance as an effective video-based app for routine fall risk assessment of older adults. The AVA will not be used as an intervention tool or a diagnosis app; instead, it is solely used as a home assessment app for fall risk. Our approach focuses on rigorous validation on older adults and studying and optimizing it for independent use by their caregivers. Our multifaceted team is a blend of AI and computer vision specialists, software engineers, UX experts centered on human-driven design, senior community wellness directors, physical therapists well-versed in fall risk assessment, and geriatricians. This project also proposes a close partnership and collaboration with Wingate University and the North Carolina Department of Health and Human Services. We will undertake a comprehensive validation and AI fine-tuning against high-end motion capture systems for accurate and reliable assessment in a lab setting (through a collaboration with Wingate University) and 100 diverse older adult participants. By bringing recent advances in large language models and self-supervised pre-trained models, we aim to make AVA a perspective-invariant assessment app to achieve reliable measurements while relaxing the setup requirements for independent at-home assessment. This project also aims to perform comprehensive user experience research and usability studies in a controlled environment, involving 50 older adults from a wide range of demographics, includ...