When working with an unidentified document, Interpret automatically tries to learn the document. (Exception: If you run Interpret using the <SYSTEM> document profile, document definitions are created in Verify.) Interpret creates a new document definition for this document. Interpret searches for the fields by first looking for the titles specified in the field profiles, and then for the corresponding field values.

If Interpret cannot learn all the fields on a document, you must complete the fields using the Verify or Optimize module.

Normally, there are complete document definitions available for most documents you process, since when you are in production you have learned the documents for all frequent senders. For these documents, Interpret uses their document definitions and processes the documents (interprets them and validates their field values). There is usually no further learning needed for these documents.

However, if more documents of the same type (invoices that have the same identifier or identifiers) come to Interpret before the document definition is complete, there are two alternatives:

• If the identifiers have previously been learned, and the new document is identified, Interpret uses the existing partially learned document definition to interpret and validate the fields.
• If the identifiers have not been learned, the new document cannot be identified. A new document definition is created, and Interpret tries to learn this definition. Later, when you use Optimize, you will indicate that the two documents belong to the same document definition and the two existing document definitions will be reduced to one.