A research team is developing an AI system for a highly specialized scientific domain where tasks are complex and require frequent adjustments. The team consists of domain experts who are not machine learning engineers. They need a method to guide the AI that allows them to easily create, test, and refine instructions in a human-readable format. Furthermore, it is crucial that they can clearly understand and troubleshoot the instructions given to the model to ensure accuracy. Computational efficiency is a secondary concern compared to adaptability and clarity. Based on these requirements, which of the two fundamental prompting approaches would be more suitable for this team? Justify your recommendation by evaluating the primary trade-offs.