We present a demo that illustrates the performance of our system to analyse and evaluate call centre conversations. Our solution can be used at different stages of the quality feedback loop. The high-level symbolic representation developed on the context-based intent recognition core module allows for detecting fine-grained reasons for quality assurance problems and going in-depth qualitative analysis of how agents and customers interact. We illustrate the evaluation and insights of real-life conversations provided by a Belgian call centre. Participants can interact with the demo by playing with call annotation, recommendations, and diverse parameters.