video1.0<iframe src="https://www.loom.com/embed/3a7b04257fea4fd49d2cae249161840b" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/3a7b04257fea4fd49d2cae249161840b-1692886541813.gif242.769In this video, I explain the divide-and-conquer approach for discovering intents. Starting from a higher level concept, I demonstrate how to break down a broad topic into more specialized intents. Using labeled data, I focus on the problem of users being charged twice for their subscription and the issue of users forgetting their username. By clustering and sorting, I identify relevant examples and label the data accordingly. This approach allows for efficient intent discovery and can be applied recursively for further breakdowns.