Sleep disorders are among the most prevalent mental disorders in the general population. They impair daily functioning, reduce quality of life and enhance health care costs for millions worldwide. To improve the quality of sleep, we propose a non-pharmacological approach based on methods and techniques derived from cognitive behaviour therapy for insomnia (CBT-I), persuasive technology and current information and communication technology. The scientific aim of this project is to enhance the understanding of personalized self-help therapy in the domain of sleep with the aid of existing (hardware) technology. We aim in particular at a generic, scientifically-proven framework that integrates persuasion strategies for sustainable behaviour change and the technology to support these strategies. As a proof of concept, a virtual coach will be modelled and implemented that enacts a first contact in a stepped care system. The resulting implementation of the model will be called the coaching engine and may be considered as a persuasive assistant that, at an appropriate time, offers and adapts sleep relevant exercises, and that gives various types of support and advice given a personal model of the coachee and its environment. In the design process of the virtual coach, an incremental approach with short evaluation cycles will be followed. In the final year, a summative user study will be performed in a Randomized Controlled Trial. Criteria for success include (but are not restricted to) improved subjective sleep quality, sleep efficiency and daytime functioning, and reduced time to fall asleep. The expected results are 1. a generic computational coaching model scientifically validated in the domain of sleep, 2. persuasive coaching strategies for exercise adherence and cognitive restructuring in computerized mobile coaching and 3. an integrative framework for symbolic interaction, sensor data, environmental changing devices and human interaction/intervention.