ESA WorldCover employs a batch processing pipeline combining rule-based and machine learning approaches, releasing a static annual layer each year, with a category system following the FAO Land Cover Classification Standard (LCCS) comprising 11 primary classes.