In the context of Exploratory Data Analysis (EDA) for a machine learning project aimed at predicting customer churn for a telecom company, the team has identified that the raw data contains customer usage patterns, service complaints, and demographic information. The project has constraints including limited computational resources and the need for the model to be interpretable by non-technical stakeholders. Given these constraints and objectives, what is the primary goal of feature engineering in this scenario? Choose the best option.