Heterogeneity will prevail in Future Networks at many technology levels, including devices, access networks, services and applications. Moreover, the exponential proliferation of mobile computing has been followed by concerns and research efforts regarding a sustainable development of wireless systems, specially related with energy efficiency. The envisioned future networking environments will be able to build a very rich source of information that can be wisely used, not only to improve users' service perception, but also to enhance network performance at different perspectives (e.g., quality of service, resource management and energy consumption). We believe that this is the main reason behind the increasingly popular concept of context-awareness. Current context-driven architectures aim to provide a platform to enable personalized services and smart networking, adapting contents and the way they are delivered to the users according to their particular context (device, available energy and preferences) and the network and environment context (location, technologies and energy consumed). However, given the proliferation of group-based communications and consequent traffic demands, we show in this paper how context information, including energy-related context, may influence content delivery to users, employing the information available to create groups of users according to their common capabilities. This way, we aim to achieve a higher level of personalization in multiparty services, improving users' quality of experience. Through a service-aware group-oriented content delivery framework, we employ algorithms for core and access network selection that are flexible enough to be guided by any constraint. The developed selection scheme uses context information from user, network and environment to decide, among the available networks, the most suitable one regarding its communications needs and optimizing network constraints, such as quality of service provided and energy required. The outcomes obtained through the simulation of the proposed framework show that context-based grouping and network selection are able to improve the users' quality of experience.