Decision Support with Artificial Intelligence has been practiced and studied for a long time. Common AI Decision Support Systems follow the same procedure: give a recommendation, maybe with explanations or other transparency features, and leave them for the human to interpret. This leaves the human decision-maker unsupported in the earlier stages of decision-making. More and more studies are converging, proposing a shift in the field of human-AI decision-making, in which recommender systems are challenged by alternative non-recommender approaches, ultimately broadening the human-AI interaction design space. This thesis builds on prior work about forward and backward reasoning decision support, where backward reasoning can be understood as recommender systems and forward reasoning as one approach to non-recommender systems. The goal of forward reasoning decision support is to place the human role at the center of decision-making, resulting in a machine-in-the-loop paradigm and a human-centered alternative