Machine learning (ML) is a branch of artificial intelligence in which software learns patterns from historical data and uses them to make predictions or decisions, rather than following only hand-written rules.
A model is trained on past examples, such as previous sales, stock levels and weather, and then applied to new data. Over time it can be retrained to keep up with changing behaviour.
Retailers use machine learning for demand forecasting, automatic reordering, personalised offers, fraud detection and price optimisation. In inventory, it can suggest what to reorder and when, based on real sales patterns.
Results depend on good, clean data and a clear question. A model is a tool to support decisions, and its output should be checked against business knowledge, particularly for new products with little history.