Advanced demand forecasting techniques that adjust to changing market conditions are essential in the fast-moving consumer goods industry. This paper presents a novel method that combines demand-sensing approaches with iTransformer models, which are known for their effectiveness in processing multivariate time series data. This methodology goes beyond typical forecasting models by combining sentiment analysis from Google Trends, Amazon reviews, and competitors’ analysis. The result is a comprehensive understanding of market behaviors and trends. The model is further improved with the incorporation of DistilBERT for sentiment analysis, which provides rich insights into customer sentiment. The results show that this model performs significantly better in terms of reliability and precision than traditional forecasting techniques, which is a considerable advancement. Using these technologies, FMCG companies can manage demand forecasts with higher precision.