This research investigates the utilization of Large Language ModelsLarge Language Models (LLM) (LLMsLarge Language Models (LLM)) to enhance Enterprise ArchitectureEnterprise architecture (EA) information retrieval within organizations. Traditional approaches often rely on the development of EA views. For instance, DoDAF–a well-known EA framework used by the U.S. Department of Defense–defines dozens of EA views representing different aspects of EAs, such as operational, system, and serviceService viewpoints. The development, access, and interpretation of these EA views present challenges for decision-makers unfamiliar with the intricacies of EA frameworks such as DoDAF or UAFUnified Architecture Framework (UAF). To address this, we developed an LLMLarge Language Models (LLM)-based tool, Oraculo. Using natural language, Oraculo enables stakeholders to query EA repositories populated with EA models. By leveraging in-context learning techniques, Oraculo adapts to the specific EA domain, translating user queries expressed in domain-specific vocabulary into the repository query language. This approach democratizes access to critical EA information, empowering decision-makers to gain insights and make informed choices regarding the evolution of their IT infrastructure without requiring in-depth knowledge of EA frameworks. Initial results demonstrate the effectiveness of Oraculo in transforming natural language queries into EA repository queries and presenting the retrieved information in a comprehensible format, facilitating improved decision-making within the EA domain.