In a recent Memgraph Community Call, Katarina from the Dev Experience team and Toni Lastre, Head of Platform at Memgraph, introduced GraphChat, an innovative feature that allows users to interact with graph databases using natural language via large language models (LLMs). This tool simplifies database interaction by converting plain English queries into Cypher queries, eliminating the need for users to understand complex query languages. During the session, Toni demonstrated GraphChat's capabilities, such as handling follow-up queries by remembering conversation context and automatically refining failed attempts, showcasing its use with datasets like the Pandora Papers and TED Talks. GraphChat is identified as a GraphRAG system, combining knowledge graphs and LLMs for enhanced data retrieval, offering an intuitive approach to querying and navigating knowledge graphs. Future enhancements include expanded context integration, error recovery, and flexibility in switching between LLM configurations, with an emphasis on enriching AI-powered search systems by integrating Memgraph with vector databases. The session highlighted how GraphChat bridges the gap between users and graph databases, providing actionable insights through conversational AI, making it a valuable tool for both managing complex datasets and exploring conversational AI applications.