The existing problems of graduates’ employment, constant changes in the labor market are a challenge for educational organizations and challenge universities to change their approaches to the organization of the educational process, forcing universities to find new tools that can involve university students in university technological entrepreneurship. For this reason, universities are developing methodological approaches and looking for formats to work with university startups that provide conditions for finding industry topics for technology startups in the agenda of industry markets. The aim of the study is to develop a methodology for identifying potentially effective student teams at the pre-experimental stage of student project selection. The authors conducted an experiment in the form of game practice with first year students of Ural Federal University, which showed that despite the random assignment of students to teams, the team with the highest cumulative frequency of meaningful words and active verbs and the most harmonious roles according to Belbin and the most cohesive work according to observers won the game. The authors conclude that game practices can identify potentially effective teams capable of generating innovative entrepreneurial ideas that can be attracted to entrepreneurship and to work in accelerators.