In modern software development issues are very important for inter-team communication and project management. Issues are used to clearly state the requirements of a change request and allow teams to plan and track their tasks. As part of an agile development process new issues are discussed during the sprint planning and reviews. However, the creation of well-structured issues with current tooling is still very time-consuming. Elements like the title and body have to be typed into the issue card and labels have to be manually selected. This makes it difficult for product owners to create digital issues during a meeting. For this reason, product owners often create handwritten notes during meetings and create issues in the issue management system afterwards. The current process results in a time and location distance between the need to create an issue and the digital documentation of this issue. This makes the process inefficient and error-prone. A product owner is effectively documenting each issue twice, once on the sheet of paper and then again in the issue management system. This thesis introduces the concept of a digital voice assistant for issue management. This system aims to automate the issue creation process and allows a product owner or developer to freely dictate an issue. Based on the spoken input a structured issue is automatically created. Elements like the assignee, labels, and priority are extracted from free text. A speech recognition system and natural language processing will be used. While modern voice assistants like Google Assistant and Amazon Alexa are increasingly common in consumer households, the underlying technology is rarely used in the enterprise context to help automate administrative tasks. The concept developed in this thesis acts as a blueprint for systems to fill out domain-specific forms, like issues or bug reports. A prototype of the system was implemented to showcase its capabilities. The system follows a four-step process. In the first step, the spoken input is transcribed using a speech recognition system. In the second step, the transcribed text is annotated with a natural language processing toolkit. Based on the annotations and transcribed text a structured issue card is filled out in the third step. The user has the option to edit and confirm the result. Finally, the resulting issue card is passed into an existing issue management system (like Github or Gropius) over an API call. To validate this solution approach an experiment was conducted. Future research possibilities and potential new use cases of the system design are presented at the end.