Socially assistive robots (SARs) have shown promise in the care of people with dementia and in mitigating behavioural and psychological symptoms of dementia. Although SARs are continually tested for efficacy, no current literature outlines a comprehensive strategy that industrial designers may employ to progress the technology of SARs. It was, therefore, essential to expand on existing literature by providing a straightforward approach to SAR design with the recommended design attributes. A systematic review was conducted to formulate recommendations for designing SARs to improve the quality of life of people with dementia. Six databases, including CINAHL, Embase, IEEE, Medline, ProQuest, and Scopus, were searched for relevant articles published between 2011 and 2022. Covidence software was used for screening, data extraction and quality testing. Of the 160 references extracted, 16 studies met the study inclusion criteria. The studies were predominately small sample sizes using various robotic platforms and technologies. Incorporating personal preferences linked to a user’s life experience and choice is a crucial ability of SARs. Natural speech communication is also an important design attribute. However, the overwhelming conclusion is that more research is needed on aesthetics, materials, and interaction capabilities. All stakeholders should be part of a holistic user-centred design process to ensure a fit-for-purpose product.