MAISTRO: Towards an Agile Methodology for AI System Development Projects

The MAISTRO methodology introduces a comprehensive and integrative, agile framework for managing Artificial Intelligence (AI) system development projects, addressing familiar challenges such as technical complexity, multidisciplinary collaboration, and ethical considerations. Designed to align technological capabilities with business objectives, MAISTRO integrates iterative practices and governance frameworks to enhance efficiency, transparency, and adaptability throughout the AI lifecycle. This methodology encompasses seven key phases, from business needs understanding to operation, ensuring continuous improvement and alignment with strategic goals. A comparative analysis highlights MAISTRO’s advantages over traditional methodologies such as CRISP-DM and OSEMN, particularly in flexibility, governance, and ethical alignment. This study applies MAISTRO in a simulated case study of the PreçoBomAquiSim supermarket, demonstrating its effectiveness in developing an AI-powered recommendation system. Results include a 20% increase in product sales and a 15% rise in average customer ticket size, highlighting the methodology’s ability to deliver measurable business value. By emphasizing iterative development, data quality, ethical governance, change and risk management, MAISTRO provides a robust approach for AI projects and suggests directions for future research across diverse industries context for facilitating large-scale adoption.