In his seminal work, Kahneman explained that human cognition results from two modes of thinking, Fast Thinking, and Slow Thinking. Fast Thinking is quick, intuitive, and non-voluntary, whereas Slow Thinking is deliberative and involves complex computations that demand concentration. The blending of fast and slow thinking is extraordinarily efficient and effortlessly optimizes our thought process. Inspired by the dichotomous model of human cognition, we propose several novel ways of operationalizing the fast-slow thinking dichotomy that can flexibly tradeoff between effectiveness and time efficiency. We present three dichotomous models, Model 1, Model 2, and Model 3, which differ based on the realizations of fast and slow thinking systems. We observe that fast thinking could be realized as a trained machine learning model or as a parsimonious CBR system that uses a subset of all the attributes. On the other hand, slow thinking can be operationalized as a CBR system that performs a well-rounded similarity estimation for prediction. Model 1 and Model 2 are realized by pairing the full-fledged CBR system with the machine learning model and parsimonious CBR system, respectively. We explore the adaptation process in CBR as a manifestation of the slow thinking process, leading to Model 3. We also present various switching strategies to efficiently switch from fast to slow thinking based on external feedback, introspection, or domain constraints. Further, for evaluating the dichotomous models, we present a footprint-based complexity measure that can quantify the tradeoff between effectiveness and efficiency in the CBR realization of the fast-slow dichotomy. Through an extensive set of experiments on real-world datasets, we show that such realizations of fast and slow thinking are useful in practice, leading to improved average response times and accuracy in decision-making tasks. We envisage that this work will pave the way for other interesting realizations not only in the context of CBR but AI in general.