Business Process Modeling enables organizations to document and improve workflows, but creating process models remains a time-intensive task requiring expertise in formal modeling languages and domain knowledge. Recent advances in Large Language Models (LLMs) have facilitated the automated generation of process models from textual descriptions, yet these LLMs often struggle with ambiguity, inconsistencies, and validation challenges. This work introduces a human-in-the-loop (HITL) approach where an LLM generates clarifying questions to resolve ambiguities, enabling human-AI collaboration through iterative refinements and structured guidance for more accurate process models. Evaluation results demonstrate that the HITL approach resolves ambiguities and improves the syntactic and semantic quality of generated process models. By bridging the gap between automation and human expertise, this approach contributes to the development of reliable and effective methodologies for Business Process Modeling.