Automotive claims prediction is a component of HyperGraf, which predicts occurrence of a claim and the claim amount for a policyholder. The underlying ML algorithms are based on variations of Regression and XG Boost using important policyholder, vehicle and GeoZone characteristics. Trained on real world claims data from an Insurance company, the algorithm considers important business factors for robustness and accuracy.