arXiv:2510.08622v2 Announce Type: replace Abstract: Software requirements are derived from a variety of elicitation techniques, many of which have a conversational nature, like interviews. However, evaluating whether those derived requirements faithfully reflect the stakeholders' needs remains a challenging manual task. In this paper, we formalize the task of aligning the transcript of an interview with a collection of requirements represented as user stories. We propose two heuristic metrics for alignment, called (i) requirements faithfulness: the proportion of stories supported by the transc
The increasing complexity and scale of software projects, coupled with the rapid advancement in AI capabilities for language processing, makes automated requirement alignment a timely and necessary development.
Automating the alignment between stakeholder interviews and derived requirements directly addresses a long-standing bottleneck in software development, promising significant improvements in efficiency and accuracy for complex systems.
The manual, error-prone task of validating software requirements against stakeholder needs can now be partially automated and quantitatively assessed, reducing human effort and improving the quality of derived requirements.
- · Software Development Teams
- · Requirements Engineers
- · AI/ML tool vendors for software engineering
- · Large enterprises with complex software systems
Reduced time and cost in the requirements engineering phase of software development.
Higher quality software produced due to more accurate and faithful requirements, leading to fewer defects and better stakeholder satisfaction.
The development of more sophisticated AI agents capable of end-to-end software development, from elicitation to implementation, further reducing the need for manual intervention.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL