The Evolution of AI-Powered Knowledge Management Systems: A Systematic Literature Review and Future Research Agenda

Abstract

Knowledge Management Systems (KMS) have evolved from rigid databases into intelligent repositories driven by Artificial Intelligence (AI). However, a systematic synthesis tracing this evolution up to Generative AI remains fragmented. Addressing this critical research gap, this Systematic Literature Review (SLR) investigates the technological evolution of AI-powered KMS, maps AI applications across the Knowledge Management Lifecycle, identifies deployment challenges, and proposes a future research agenda. Following PRISMA 2020 guidelines, a total of 105 peer-reviewed articles published between 2018 and 2026 were retrieved from Scopus, IEEE Xplore, and Web of Science for qualitative and quantitative synthesis. The review identifies three evolutionary eras, culminating in Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). Quantitative analysis reveals that AI integration is heavily concentrated in knowledge retrieval (representing 45% of analyzed studies) and creation. However, critical deployment barriers persist, particularly model hallucinations, enterprise data privacy, high computational costs, and user adoption resistance. This study distinguishes itself from previous reviews by providing a structured taxonomy of AI paradigms in KMS, offering strategic blueprints for enterprise deployment, and outlining future research trajectories focusing on Graph-RAG, multi- modal agents, edge-deployed local models, and AI governance frameworks.