Forecasting the demand for new fashion products in the fast fashion industry is a com-

plex task due to its dynamic nature, short product life cycles, and limited historical data.

Traditional forecasting models often fail, leading to inefficiencies such as overproduc-

tion or underproduction. This paper reviews key challenges and explores innovative

machine learning (ML) and artificial intelligence (AI)-based models to improve fore-

cast accuracy. We propose a hybrid AI-driven approach that integrates structured and

unstructured data sources, real-time monitoring, and ensemble models to address fore-

cast limitations in the fast fashion industry.