In order to build and maintain customer trust, financial institutions are investing millions of dollars in preserving privacy while experimenting with exponential technology to bring best value to their customers. With rapid advancements in artificial intelligence on all fronts, however, there is a dire need in the financial sector to balance customer privacy with the use of transaction data in training robust machine models. This white paper provides a rationale focused on data privacy and also includes suggested algorithmic recipes related to how transaction data could completely eradicate the use of real customer data. A distinction is drawn between data simulation and synthesis. Two discussed synthesis techniques include a bootstrapping framework featuring dynamic time warping and a hidden Markov model fitted with a stochastic gradient version of the Markov chain Monte Carlo method.