The widespread deployment of intelligent systems in critical sectors, such as healthcare, justice, and finance has underscored the urgent need for transparent and trustworthy in the answers provided. Accordingly, explainability constitutes a key requirement to foster understanding, trust, and accountability in automated decision-making processes where AI in general, and Large Language Models (LLMs) in particular, are involved. This paper introduces a new explainability framework for symbolic systems, centered on an application pipeline aimed at constructing a semantic knowledge graph from unstructured text. In our approach, explainability for the symbolic entity recognition component is achieved through an event-driven system, where each operation is logged atomically in a journal, capturing events that trace object creation, lineage, and purpose. Alongside the journal, two additional components support the architecture: the Entity Storage, which maintains versioned records of entities, and the Explainability Module, which interprets and explains the logic behind each entity's representation. The proposed framework offers a high degree of detail and customizability and is currently being tested within a medical application designed for general practitioners. In this context, the ability to provide transparent justifications for specific recommendations is essential to ensuring reliability for both physicians and patients. Furthermore, it contributes to strengthening the credibility and trustworthiness of AI-based tools, which are expected to assume an increasingly central role in this domain.