In agile software development, maintaining high-quality user stories is essential yet challenging. This study investigates the potential of large language models (LLMs) to automatically enhance the user story quality for agile teams at a case company. We proposed a reference model for an Autonomous LLM-based Agent System (ALAS) and implemented it in a mobile delivery project. The effectiveness of the ALAS in improving user stories was assessed through a questionnaire completed by 12 participants from six agile teams. Our findings highlight that LLMs, leveraging their natural language processing capabilities, have the potential to significantly improve user story quality. This research not only contributes to the discourse on AI's role in agile software development but also demonstrates a practical example of AI's ability to revolutionize industry practices, providing a proof-of-concept for integrating AI technologies into requirements engineering tasks.