International depository Clearstream is collaborating with the University of Luxembourg to investigate the use of artificial intelligence in determining user expectations and functional specifications for new IT systems.
The four-year project will examine the potential of AI-based systems and natural language processing in producing outcomes more aligned with business requirements when developing new software.
Clearstream says the initiative, run in collaboration with business analyst firm escent, will seek to automate the laborious analysis tasks present in Clearstream’s IT system requirements engineering practices.
This will range from reconciling requirement models with accompanying natural-language content, to deriving system acceptance criteria from requirements.
Prof. Lionel Briand, vice director of the University's Interdisciplinary Centre for Security, Reliability and Trust, says: “Nearly 50% of budget overruns in IT projects are caused by inadequate requirements which ripple through system design and deployment. Together with our partners, we will work at the cutting edge of new technology to introduce automation to the requirements engineering process - reducing incompleteness, inconsistency, and ambiguity to a minimum - and facilitating the analysis of system compliance.”