Programme Launch Offer: Save 20% - Book Now

Track Talk, T14

AI-Powered Learning in QA and Beyond

Nils Röttger

14:15 - 15:00 CEST, Tuesday 16th June

Today’s software development is constantly evolving, and we testers must continuously deal with new challenges: DevOps, security testing, and AI testing – the required knowledge must be acquired quickly and efficiently. In addition, there are the technical requirements in our own domain and in our customers’ domains within the respective projects. How can we learn efficiently? How can we acquire the specific domain knowledge of our customers?

In this talk, I will demonstrate how language models such as Mistral or ChatGPT can function as coaches, trainers, or research tools for agile learning in quality assurance – individually, practice-oriented, and on demand.

In several experiments, we have developed and followed learning paths for various QA roles – including test automation engineer, test automation architect, requirements engineer, and project manager. For the project manager role, for example, we used AI to prepare for a conflict conversation in a chat session, with the AI simulating a frustrated employee.

However, the talk will focus on the new role of the AI Quality Engineer, which emerges as an independent specialization in response to the increasing integration of AI in software projects. We have also developed learning paths for technical domain knowledge, for example, for understanding a customer’s CT devices. We used AI models both for developing the learning paths and for the learning process itself. The language model was strategically deployed to identify knowledge gaps, prepare learning content, and respond interactively to questions – truly in the spirit of an agile learning coach.

The talk demonstrates the diverse possibilities that AI-supported learning assistants offer for quality assurance – ranging from personalized learning methods to accompanying complete learning paths within organizations. It also becomes clear how crucial the right interaction with the language model is: good prompts become the key to effective learning. The talk not only provides technical insights but also demonstrates how these approaches can be meaningfully integrated into existing training concepts, such as agile learning.