Conceptual approaches for the use of artificial intelligence algorithms in the functional screening of secondary metabolites following the bottom-up approach. A library of secondary metabolites obtained from metabolomic studies can be screening for biological activity using their chemical structures. In a machine-learning approach, an artificial intelligence system will be previously trained with a reference dataset where the important chemical features will be manually curated. In a deep-learning approach, the artificial intelligence will extract the functional features and classify the database of chemical compounds with no human intervention.