Enterprise AI Is a Platform Problem: GKS-5, A Reference Architecture for Governed Knowledge Systems

GKS-5: A Reference Architecture for Governed Knowledge Systems presents an enterprise architecture for governing AI-enabled knowledge systems as they move from isolated experimentation toward operational use. The paper argues that enterprise AI should be treated as a platform and governance problem rather than only as a model-selection problem. It introduces a structured reference architecture for controlling how enterprise knowledge is accessed, transformed, evaluated, governed, and consumed by AI-enabled applications and agents. GKS-5 focuses on the architectural separation of knowledge, governance, policy, assurance, identity, provenance, and execution concerns. The objective is to provide organizations with a practical foundation for building AI systems that remain auditable, policy-aware, explainable, and manageable across heterogeneous enterprise environments. The work is intended for enterprise architects, AI platform teams, security and governance practitioners, researchers, and organizations designing trusted AI and agentic systems.