AI-Kit is a concept in development at Bureau Moeilijke Dingen that allows users to envision, design, and develop their own functionality using Machine Learning. During my graduation project, the opportunities and challenges of implementing this system in Smart Homes have been explored. The design scope and theoretical background has been defined, and a descriptive model of the system has been created to obtain an overview of the required devices. Using the model and theory, different required interactions (e.g. system feedback, output control, and exception flagging) were defined. The (dis-)advantages of potential interaction locations were hands-on explored (input/sensors, context/feedback, output/IoT Devices). Concluding to the work, a demonstrator has been developed that incorporates the findings of the process allowing the design requirements for the next semester to be clearly defined and communicated.