Extracting structured data from documents by following a user-defined template or schema that specifies what fields to extract.