Answers are grounded in this paper and cite their sources.
Organizations have long recognized that their most strategically valuable knowledge is precisely the kind that resists formalization: tacit knowledge, embedded in expert judgment, professional routines, and situated experience, has historically eluded the capture mechanisms that knowledge management systems were designed to operationalize. Classical elicitation methods addressed this challenge at the cost of scale, while digital repositories excelled at scale at the cost of depth, a tension that has defined the central paradox of enterprise knowledge management for decades. The advent of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) architectures has been framed in the literature as a qualitatively different paradigm, offering systems that are theoretically capable of engaging knowledge holders in flexible, context-sensitive dialogue while simultaneously organizing and recombining heterogeneous knowledge sources at organizational scale, suggesting the possibility of a new epistemological infrastructure for the firm. Grounding the analysis in the SECI model, this paper provides a deep analysis on how naive, advanced, and agentic RAG architectures map onto each phase of organizational knowledge creation, critically assesses the degree to which AI-mediated approaches can complement classical elicitation methods, and surveys enterprise applications across industrial, medical, legal, and social service domains. Structural limitations, including hallucination risk, graph construction complexity, and evaluation framework gaps, are identified, and a research agenda at the intersection of generative AI and organizational epistemology is proposed.
No takes yet. Share an insight, caveat, or question.
Urmeneta et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: