We are working on a smart mobile service to help people take control over their diabetes. Artificial intelligence algorithms will help to recognize the physical state of the person and provide personalized feedback. The solution is data-focused - we rely on collecting and processing health-related data such as glucose levels, types of insulin, food, physical activity, stress, etc. Information will be collected from low-cost sensors and mobile devices. The service will run in the background, helping with day-to-day control of the disease. It will interact with the user and offer suggestions. It will make predictions about the state and intentions of the user. It will be context-aware and tailored to the user, thanks to learning his behavior. Data analytics will help the patient to discover new patterns in his diabetes and manage the disease better.