Abstract: A process for recommending items comprises receiving a first input from a user via a user interface and modifying the user interface based on the first input. A second input is received from the user via the user interface. The first and second inputs are used as data for running a set of primary models and a set of secondary models, where each item that the process may recommend corresponds to one and only one primary model and one and only one secondary model. The primary models calculate a confidence level that the user should be recommended the item associated with the primary model, based on the first and second inputs. Items associated with primary models where the confidence level is above a threshold are recommended to the user. For each recommended item, a parameter is generated by the associated secondary model.