Abstracts

Customers are interested in exciting features and functionalities. What they have the right to expect, they take for granted and thus is less interesting.
How to find test cases that address customer needs, and give your product a competitive advantage? Quality Function Deployment (QFD) can be used to generate those test cases automatically for features that excite customers. This method relies on a close collaboration between test department and product marketing; thus, it is not easily implementable in an average industrial environment. However, those who master this challenge get high rewards.

Given the complexity of many modern products, a human-guided, manual QFD approach often is inappropriate because it is too complex. Artificial Intelligence (AI) has the tools that are needed to apply the QFD methods even to very large and complex SW products.

New methods are needed for interoperability testing digital systems, according to the new Interoperable Europe Act. It is inconceivable how this should be mastered with traditional testing methods alone. Automated testing based on customer, interoperability, and regulatory needs is the way to go, adopting methods from QFD and AI.
While QFD has a at least a 50-year-old history, and AI 80 years, using QFD methods for AI is a novel approach. QFD adds explainability and better human-machine interaction to AI. Technically, Large Language Models (LLM) are inflated QFD matrices – however, as the matrices inflate from 12 x 20 to 40’000 x 70’000, this leads to completely different new applications, while the approach remains the same.