A financial institution is developing an AI model to assess customer credit risk. Regulatory auditors require the model's decision-making process to be understandable. The data science team is debating whether to use a complex deep learning model or a simpler logistic regression model. What is the best approach to meet both regulatory and business needs?