Early Bird Discount Ends June 30! Get a special discount on Combo tickets before prices go up. Secure yours now — Click here.
From band-aids to self-healing: AI-powered bug management
Track Talk I November 14 I Track 1 I 2:45 PM – 3:25 PM
Description
Bug management isn’t just about tracking issues—it’s about identifying patterns, preventing recurring defects, and improving software quality. At Choco, an AI-first company fighting food waste globally, bug triage was a major bottleneck. Developers wasted hours searching for past fixes, sifting through Jira tickets, and retracing debugging steps. What if AI could change that? An AI-powered bug assistant transformed bug triage—recommending fixes, learning from feedback, and improving over time. But it wasn’t perfect. The initial RAG solution struggled with incorrect or outdated recommendations, and keeping the knowledge base relevant was a challenge. Through iteration, the system evolved to automatically update itself, learn from user feedback, and refine recommendations. This session focuses on practical lessons from building an AI-driven knowledge system, covering:
* The initial RAG solution – Using closed Jira tickets as a structured knowledge base to generate AI-driven bug recommendations.
* Keeping knowledge fresh – Preventing outdated solutions with automated knowledge base updates, as bugs and features evolve rapidly.
* Introducing the feedback loop – Tracking user feedback on AI recommendations to improve accuracy.
* Self-Improving RAG – Turning incorrect AI responses into learning opportunities by generating FAQ entries in Notion and automatically updating the vector database.
* From low-code to full control – Moving from n8n low code automation to LangGraphJS for better debugging, multi-agent workflows, and long-term scalability.
* What’s next? – Corrective RAG for smarter error detection and agentic AI that continuously refines itself.
This session is for testers, developers, and AI enthusiasts looking to integrate AI into testing and quality workflows. Whether you’re exploring AI-powered knowledge bases or actively building them, you’ll leave with concrete strategies to make them accurate, scalable, and self-improving. Join me to see how AI-driven bug insights shift teams from reactive firefighting to proactive quality assurance—evolving from patching issues to building self-healing systems.
Key-takeaways
- AI-powered bug Management that works – Retrieval-Augmented Generation (RAG) turns bug tracking into an active problem-solving tool, helping developers find and apply fixes faster instead of just searching through past issues.
- Feedback loops drive AI accuracy – Without continuous learning, AI-generated recommendations lose relevance. A structured feedback system ensures AI evolves, improving accuracy and delivering reliable, context-aware solutions over time.
- The tech & future vision of RAG solutions – From low-code (n8n) to full control with LangGraphJS, multi-agent systems can self-correct, route knowledge intelligently, and refine responses—shifting AI from static retrieval to adaptive, scalable intelligence.
About the Speaker
Anaïs van Asselt is a quality coach and test automation enthusiast with over 10 years of experience in web and backend testing across various organisations. She works closely with developers, applying a context-driven approach to sustainable test automation. Her passion for sustainability—both in QA and beyond— led her to move to Berlin in 2022 to join Choco, a startup on a mission to reduce global food waste. As one of two quality engineers at Choco, she works with product teams to embed quality-first principles within the SDLC, standardise test automation, and integrate QA solutions into CI/CD pipelines. In the past year, she has been diving deep into AI and automation, exploring innovative ways to integrate AI-driven solutions into testing, and quality processes.
