Your team trained a BigQuery ML model to predict loan default risk, and you are now running inferences on new data. To ensure that the model predictions are interpretable and useful to the finance team, which two actions should you consider? (Select two)