Assistive technology (AT) products and services are increasingly used to support persons with dementia (PwD) and their caregivers, in terms of healthcare, safety, autonomy, leisure and social participation. Studies conducted in this area have tended to focus on usability engineering and AT acceptance, rather than on AT provision and follow-up plans. In other fields of disability, efforts have been made to integrate AT modeling frameworks into delivery practices, including the selection and assessment of AT over time. In the context of dementia, probably because of the relative novelty of the use of AT, only a few works have stressed the need for a comprehensive framework to guide users, practitioners and product developers in decision-making regarding the conception, evaluation and provision of AT. In this paper we provide preliminary guidance for the definition of such a framework. For doing so, first we review two existing AT models, chosen because of their applicability in the field of dementia: Hersh & Johnson’s “Comprehensive Assistive Technology” model and Scherer and colleagues' “ICF core set for Matching Older Adult with Dementia and Technology”. Then we discuss some implications of the use of AT models and frameworks for clinical practice, specifically their incorporation within the integrated care systems increasingly adopted worldwide. Subsequently, we propose a set of key factors that should be considered for building tools to support AT design, provision and assessment in the context of dementia: the progressive nature of the disease, the clinical heterogeneity observed among PwD and the subsequent need for personalized care plans, the dynamics of function allocation between PwD, AT and caregivers, and the role of fluctuating symptoms and preserved abilities in this population. Finally, we suggest some directions for further research in this field.