Product Backlog

The dataset includes 22 projects and 1680 user stories, with the aim of classifying these stories into those suitable for AI implementation and those not recommended for AI implementation. The labeling was done in a group, reaching a consensus on each user story in each project, determining whether it is susceptible to being developed with AI. Thus, each user story was evaluated and assigned a value of 1 if it was considered suitable for AI implementation (this label was named AI), and a value of 0 if it was not (this label was named not-AI).