Microsoft's Azure Content Understanding prebuilt analyzers extract invoice and tax form fields

Microsoft says its Azure Content Understanding prebuilt analyzers extract structured fields from documents such as invoices, tax forms, and identity documents, using a maintained schema plus built-in knowledge of each document family.

For supported fields, the analyzers return normalized typed values, confidence scores, and grounding that links each result to its source span, page, and coordinates.

Microsoft says the approach distinguishes similar amounts, such as a requested amount versus another total, and handles locale-aware number and date formats.