Considerations
- All data items requiring extraction to be clearly legible and in English.
- Documents or mail content must include and clearly define any required keys and values.
- Non-standard layouts may not be possible for OCR to process accurately with high confidence. Page formats with multiple columns, nested, adjacent or overlapping tables, forms, or key pairs, and so on, must be reviewed during implementation.
- Handwritten, manually edited documents and any other manually annotated documents can require manual processing.
- Documents in the form of Pdf are only supported as an email attachment.