Architectural Artificial Intelligence Design Patterns for Knowledge Management Processes

ABSTRACT This paper explores the emerging architectural artificial intelligence (AI) design patterns that support today’s knowledge management processes. It summarises first various AI disciplines, then reviews related works aiming at providing design patterns. It then presents innovative solutions based on real-world use cases developed with our students and discusses the AI-based systems lifecycle as well as common underlying patterns such as agents, RAG or GraphRAG. For each pattern, we describe the problem it addresses, the solution it proposes, as well as its main benefits and illustrate with business use cases. Design patterns are a well-known technique for improving software engineering and the quality of software. As AI-based systems evolve, it becomes increasingly important to document emerging AI design patterns.