Annotations now follow your document's layout
Extracted fields are now ordered to match the layout of the source document, both in the validation view and in the output. This makes it faster to check extracted data against the original document, and gives you a predictable field order to build integrations against.