Improving the Quality of your Requirements with Artificial Intelligence (AI)

Guidelines, rulesets, trainings, boilerplates, criteria, checklists, approaches, … to improve requirements quality have long been around – INCOSE, IREB, EARS, … State of the art LLMS like GPT-4o are already ‘pre-trained’ on all of them.

Our combined 50+ years of customer experiences show that engineers still struggle tremendously to come up with (new) requirements of good (or at least reasonable) quality; and that tons of existing bad quality requirements are additionally making the situation much worse for basically each and every company / project. The huge negative impact on project success versus failure, delays, costs, … is undisputed. This presentation addresses both challenges starting with re-going through the main aspects of good quality versus bad quality requirements.

Transparently Integrating chatbots such as Chat-GPT, while customizing them by ‘attaching’ requirements engineering and company specific expertise, allows engineers to fully benefit from AI in tools like Visure Requirements ALM:
1. Spotting weaknesses, like:
a. conflicts and inconsistencies;
b. duplications;
c. ambiguity;
d. bad structure;
e. not being testable;
f. containing multiple imperatives;
g. …
2. Getting improvement proposals;
3. Apply re-writing proposals;
4. …

We encourage participants to submit their candidate requirements (specifications) to make it their story during a short demo.