Ontology population consists of automatically identifying, extracting and integrating information from different sources to instantiate the classes and properties of an ontology, thereby building a knowledge graph of a domain. In this article, we compare two ontology population techniques from texts: KOnPoTe, a semantic approach that relies on textual analysis and domain knowledge analysis, and a second Generative AI approach using the Claude model, which is based on a Large Language Model (LLM). Experiments conducted on French real estate advertisements are presented. The advantages and limits of both approaches are discussed.