A data scientist trains an XGBoost classifier for customer churn prediction. After training, both the training accuracy (72%) and validation accuracy (71%) are low and nearly identical. Plotting the learning curves shows both curves plateau at a high error level and converge close together. Which conclusion and remedy are most appropriate?