Machine learning-based adaptive learning environments adapt in real time to users and their learning state. Information processing of adaptive learning environments can be based on different models. This research work deals with the practical implementation of a three-model architecture in an AI-based learning environment. It investigates the didactic requirements for every component of the model, the techniques that can be used to design and implement the components in an adaptive learning environment, and the challenges that arise. The tools and didactical concepts to be selected represent the current state of development. The focus of this paper is to demonstrate simple strategies for conceptualizing adaptive learning environments so that future creators of these environments, including media didacticians, educators, and developers of modern educational technologies, can independently design their own adaptive learning environments.