Most of the world’s population is using smartphones and other wearables which produce large amounts of raw data that can be used by AI to monitor behaviour patterns, activities, habits, detect and predict behavioural change, anomalies, etc. If this information is enriched with physiological data it can be analysed in such detail to not only provide suggestions on how to improve one’s lifestyle but also to support efficient self-management of various diseases. Including such services into connected health infrastructure can also provide the information about the state and history of the patient’s parameters and wellbeing upon which they can react and make prompt decisions. We will present our experience and use-cases in terms of AI research and applications in domains of lifestyle monitoring, e.g., physical activity, mental stress, and environment as well as recognising and predicting health conditions, e.g., discovering common medical conditions, predicting glycaemias in diabetes, predicting hospitalisations in congestive heart failure, etc.