Publications in recent years describe a strong use of wearable systems (especially smartphones) in the eld of Human Activity Recognition (HAR). The smartphone can be considered both as a set of sensors and as an interface for the user. The European Horizon 2020 e-VITA project aims to promote active and healthy aging in senior populations. To this end, and thanks to Information and Communication Technologies (ICT), the current work is focused on the development of a virtual assistant to help the user in his daily tasks. One of the ICT used in this project is the smartphone. Indeed, the sensors of the smartphone make it possible to measure and reconstitute a part of the user's context, thanks to the communication with a remote server. Moreover, this server is able to o er di erent services (including the fusion of di erent modalities) in order to reconstitute the context. The fusion of multimodal data implies in its turn di erent machine learning modules. Finally, in order to e ciently restore the context to the user and to dialogue with him, a chatbot module under the chatGPT model developed by openAI is integrated. The communication with the user will then be established in a textual or vocal way.