In the advent of miniaturized mobile devices, on a growing scale enriched in computational, storage and wireless communication resources, the need for delivery of highly-personalized, intelligent mobile applications anytime-anywhere-anyhow is becoming a must for any successful service provider. Mobile service users expect high personalization, context-awareness and intuitive interfaces based on their needs at hand; yet, current challenge lies in recognition of user's context and needs. In our approach, a service collects data using an array of diverse ubiquitous sensors embedded in user's devices and environment, and learns a model of the ‘user's world'; his patterns of habits, usual contexts and needs. Necessarily, the service learns on a continuous basis; along the user's mobility in his spatial-temporal, personal, professional, social and technological spaces. This process we denote a mobile cognition. Based on the learned ‘user's world', a cognitive mobility process, proactively and intelligently adapts the service delivery to the user's needs. We demonstrate the efficiency and effectiveness of such an intelligent service delivery in the Quality of Service (QoS)-information System (QoSIS) case study for a mobile healthcare (mhealth) service, i.e., where the user's health state is a human aspect of the service's intelligence and becomes the user's context. We propose the use of the concepts of mobile cognition and cognitive mobility for delivering intelligent services at large.