Visitors are requested to answer questions by speaking them into a microphone. The answers are then transcribed into text, tagged and classified using large language models. When submitting an answer, it can be traced on the projection as it makes its way into the data graph.