How to group document modules

This guide outlines the best practice for grouping documents modules to ensure the needed results and prevent errors.

The v2 document modules depend on each other, so they must run in sequence. Each module uses the output of the one before it.

Required module order

1. Document Classification
2. Document Extraction
3. Document Authentication

Document Extraction depends on Document Classification, and Document Authentication depends on both. Because Document Authentication checks the extracted fields against the document image, it can’t run before Document Extraction or in parallel with it.

Document Classification produces the anchor image, so a Facematch Verification module can run in parallel with Document Extraction and Document Authentication. To learn more, refer to the Combine document and biometric modules guide.

Best practices

• Start with Document Classification. The module scans the submitted document to determine its type and validity.
• Keep Document Extraction before Document Authentication. Authentication depends on the extracted fields, so it can’t run first or in parallel with extraction.
• Evaluate after Document Classification where you want to end unclassified journeys early.

The required order of the document modules is fixed, but the evaluation steps around them are not. Choose the evaluation and routing that best fits your use case.

Read the Set up evaluation documentation to learn more about evaluations.