The article presents Gaia, a methodology for agent-oriented analysis and design. Gaia is both general and comprehensive, applicable to a wide range of multi-agent systems and addressing both macro-level (societal) and micro-level (agent) aspects. It views multi-agent systems as computational organizations consisting of interacting roles. The methodology is illustrated through a case study on an agent-based business process management system. The article discusses the need for software engineering techniques tailored to agents, as existing methods like object-oriented analysis and design are inadequate for capturing agents' flexible, autonomous problem-solving behavior, rich interactions, and complex organizational structures. Gaia is introduced as a methodology specifically tailored for agent-based systems. The article is structured with an introduction, domain characteristics, a conceptual framework, agent-based analysis, design, a case study, related work, and conclusions. The domain characteristics of systems suitable for Gaia include coarse-grained agents using significant computational resources, systems aiming to maximize global quality measures, heterogeneous agents implemented with different programming languages and techniques, static system organization and agent abilities, and a small number of agent types. Gaia addresses both macro and micro aspects of design and is neutral with respect to the target domain and agent architecture. It represents an advance over previous agent-oriented methodologies. The methodology allows analysts to systematically move from requirements to a detailed design that can be implemented directly. The requirements capture phase is independent of the analysis and design paradigm.The article presents Gaia, a methodology for agent-oriented analysis and design. Gaia is both general and comprehensive, applicable to a wide range of multi-agent systems and addressing both macro-level (societal) and micro-level (agent) aspects. It views multi-agent systems as computational organizations consisting of interacting roles. The methodology is illustrated through a case study on an agent-based business process management system. The article discusses the need for software engineering techniques tailored to agents, as existing methods like object-oriented analysis and design are inadequate for capturing agents' flexible, autonomous problem-solving behavior, rich interactions, and complex organizational structures. Gaia is introduced as a methodology specifically tailored for agent-based systems. The article is structured with an introduction, domain characteristics, a conceptual framework, agent-based analysis, design, a case study, related work, and conclusions. The domain characteristics of systems suitable for Gaia include coarse-grained agents using significant computational resources, systems aiming to maximize global quality measures, heterogeneous agents implemented with different programming languages and techniques, static system organization and agent abilities, and a small number of agent types. Gaia addresses both macro and micro aspects of design and is neutral with respect to the target domain and agent architecture. It represents an advance over previous agent-oriented methodologies. The methodology allows analysts to systematically move from requirements to a detailed design that can be implemented directly. The requirements capture phase is independent of the analysis and design paradigm.