Hazardous alpine mass movements like avalanches, debris flows and rock falls generate seismic signals, which can be detected with conventional earthquake seismometers. These signals are often emergent, weak and depend on the kind of mass movement as well as the local geology at the source, sensor and in between. This makes automatic detection of mass movement seismograms difficult. Here we investigate the use of intelligent algorithms to scan realtime data streams for mass movement signals. Given the rapidly growing data volume of national and international seismic data repositories and the increasing sensor coverage even in remote terrain, we aim to significantly enhance our ability to monitor and warn against Alpine natural hazards.