Abstract: A system and method that allow an institution's customer's anticipated data to be stored in a cache in advance of a request for that data, using predictive models based upon analyses of each specific customer's historical patterns. For example, when a customer logs on to an institution's server, a machine learning application on that server may predict the type or category of data that the customer may intend to retrieve, based upon his or her previous patterns when logging on. Typical categories of data may include, for example, account balances, status of outstanding loans, credit card data, and insurance information. By populating the cache with such anticipated data, that data may be more quickly accessed by the customer.