The evolution of computing technology has created several new avenues for artificial intelligence to become part of various industries. The most important drawback in AI systems, especially for the financial sector, is the difficulty in interpreting complex and diverse requirements. This study introduces the application of scenario-based requirements engineering (SBRE) offering a comprehensive methodology that addresses this challenge for developing AI-based credit decision systems. SBRE offers a novel approach in handling the complexity of AI-based systems that is adaptive to changing data and operating conditions. Unlike traditional approaches, this study integrates dynamic scenarios for validation and verification of requirements, which ultimately improves the credit model accuracy, reduces risks, and ensures requirements are met. This study makes a significant contribution in expanding the scope of requirements engineering for AI-based systems and paves the way for further exploration in other sectors. © 2025 IEEE.