Large Language Models have enabled significant improvements across a wide range of Natural Language Processing tasks, making them the interface of choice for many novel data exploration systems. However, LLMs also have known shortcomings, including hallucinations and cut-off dates. In this talk, I will present the important role of Knowledge Graphs in complementing LLMs with accurate and up-to-date information. The SIB Swiss Institute of Bioinformatics has a long-standing expertise in curating and maintaining high-quality, interoperable knowledge graphs across diverse bioinformatics disciplines. I will discuss how LLMs provide today an important opportunity to democratize access to these scientific data for a wider audience of researchers, if used appropriately. Finally, I will demonstrate a concrete application leveraging LLMs for Knowledge Graph Question Answering.