Rule-based extraction relies on predefined rules, regular expressions, and templates, making it well suited to known documents with fixed layouts. Model-based extraction uses trained machine-learning models—including prebuilt and custom models—to infer target fields from structured or semi-structured documents whose layouts may vary. UiPath documents its Form Extractor as template- and rule-based for non-variable formats, and its Machine Learning Extractor as suitable for structured or semi-structured documents with widely varying provider layouts.
Community Discussion