Machine learning and AI systems are most commonly used in the finance business function in order to forecast future expenditure for budget allocation and prediction. Fast Fourier transformations are able to generate stable approximations of even the most complex seasonality patterns while recurrent neural networks capture non-linear asymmetric cyclic patterns. However, with the advancement of sophisticated algorithms and computational power, anomaly detection systems are able to traverse internal financial databases, proactively searching for fraudulent or abnormal financial behaviour Рautomatically notifying the financial department when abnormalities are discovered.