Automating Non-Functional Requirements with AI
Using LLMs to enhance software quality from the start
This research introduces a framework that uses Large Language Models to automatically generate non-functional requirements from functional specifications, addressing a critical gap in software development practices.
- Leverages custom prompting techniques in a Deno-based pipeline to identify quality-driven NFRs
- Helps requirements engineers avoid overlooking critical quality attributes like security, performance, and reliability
- Prevents costly rework by incorporating quality considerations early in the development lifecycle
- Demonstrates how AI can augment engineering expertise in requirements elicitation
For engineering teams, this approach offers a practical solution to the persistent challenge of comprehensive requirements gathering, potentially improving software quality and reducing technical debt.