Decision making is one of the most interesting areas of scientific research in social sciences. Behavioral economics has shown us that, contrary to previously held opinion, individuals are not entirely rational agents and fall under the influence of cognitive biases that result in non-optimal decisions and outcomes. The problems described occur not only in traditional situations such as shopping, but also in all important life decisions such as those that students have to make during their education. As it is extremely difficult to make decisions whose consequences last a lifetime, people consider a number of factors in such decisions. Furthermore, it has been shown that students of certain academic branches are more inclined to look at external factors than others. This puts them in danger of falling under the influence of the herd mentality, that is, of following the decisions of the majority, even if they are not the best decisions for themselves. Research has shown that such an effect can be avoided if students do not follow the majority, but individuals who are most similar to themselves. With this idea in mind, the author developed a special application with a recommendation engine based on finding the most similar alumni to the student (through personality, career interests and study preferences) and using their experiences to recommend to the student which specialties might be best for him. This paper presents the theoretical background of the problem of choosing a field within the context of education, shows the current situation in which students of the Faculty of Economics and Business in Zagreb choose a field, describes the development and presents the finished EFZG smjer web app - which incorporates the proposed recommendation system.