My Product Owner is a Robot: Using ChatGPT to Create User Stories

This analysis evaluates the efficacy of using generative AI, specifically ChatGPT, to assist in the creation of user stories for game development. The experiment focuses on deconstructing the Golden Ticket feature from the mobile game Candy Crush Saga. By utilizing a public wiki description as a foundational design brief, the study tests whether AI can effectively translate high-level feature concepts into granular, actionable user stories suitable for agile development workflows.

The findings indicate that AI significantly accelerates the documentation process, reducing a task that typically requires hours of manual effort to less than ten minutes. The generated output demonstrated a high degree of structural competence, successfully incorporating the benefit-driven format—a common point of failure in manual story writing—and adhering to the INVEST criteria (Independent, Negotiable, Valuable, Estimable, Small, and Testable). Furthermore, the AI proved capable of decomposing complex features into both player-facing stories and backend operational requirements, even generating a structured user story map backbone.

While the generated stories are not intended for immediate production integration without human oversight, the results suggest that AI serves as a powerful catalyst for project management. It functions effectively as a "rough sketch" tool, providing a solid foundation that producers and product managers can refine. The study concludes that while AI lacks the deep organizational context required to identify nuanced edge cases, it is a highly efficient instrument for enhancing the speed, depth, and clarity of the user story development process in game production.