ReasonCast Framework for Agentic Demand Forecasting via Semantic Interventions

ReasonCast uses an agentic framework to convert event context into structured semantic interventions rather than generic text embeddings. This method distinguishes between forecast-relevant semantic effects and irrelevant noise to improve time-series predictions.