Recommendation as a service in smart cities: A scoping review and framework for adaptive quality of experience

Recommender systems are central to digital services, yet their development and integration remain costly and complex. Recommendation as a Service (RaaS) addresses these challenges by providing scalable, ready-to-integrate personalization solutions. Despite this potential, research on RaaS remains fragmented, with limited attention to its role within smart city ecosystems. Smart cities rely on digital infrastructures to optimize resources and improve citizen well-being. Embedding RaaS in this context can enhance the Quality of Experience (QoE) by tailoring services to user needs, preferences, and situational factors. However, most studies focus on isolated domains or narrow indicators, leaving broader system-level interactions underexplored. This paper presents a scoping review of RaaS with an emphasis on QoE in smart cities. A survey of recent literature is conducted to identify gaps, challenges, and opportunities. Building on this, we propose a QoE-enabled RaaS architecture featuring a dynamic workflow that adapts recommendations through continuous monitoring of human, system, and contextual factors. We illustrate its applicability through representative Connected and Autonomous Electric Vehicle (CAEV)-oriented deployment scenarios. The findings of this review and the proposed reference architecture aim to guide, support, and accelerate the development of adaptive, trustworthy, and citizen-centric recommendation services.