Any group housing of livestock carries the potential for competition over scarce resources and, as a result, an increased risk of injury. In the case of horned goats, this can have particularly serious consequences. Close monitoring of animal behavior is therefore essential. The goal of the VerZi project is to automate herd monitoring using artificial intelligence. By quantifying behavioral changes, it becomes possible to objectively evaluate the effectiveness of measures designed to reduce the risk of injury without the injuries actually having to occur. In addition to computer vision methods, bioacoustic methods are also used.