This workshop focuses on designing structured workflows and Quality Gates for AI-powered QA agents, with a strong emphasis on ensuring quality. Rather than building the agents themselves, participants will focus on defining how these agents should operate within a controlled workflow, including how their outputs are validated at each step.

The main objective is to ensure the reliability and consistency of AI agents by defining minimum acceptable quality for outputs, creating deterministic validation steps (Quality Gates) across the workflow, and identifying when and where adjustments are needed. We approach this as a pipeline of tasks, where each step is validated to guarantee quality before moving forward.

In addition, we introduce observability practices to monitor agent behavior, detect failure points, and enable continuous improvement.