Data extraction
- Olli Järvinen
The agreed index fields that can be found in the document will be indexed. The indexed data is still manually validated to ensure high quality. If the field to be indexed is empty, the index field is marked with a predefined string.
Line items are extracted based on line headings. The information should be in the lines correctly titled and in an understandable form. References to header level information are not taken into account.
Note that base offer only includes header level extraction. For line level extraction, please contact Staria sales.
Data interpretation supports multiple languages, which are listed in Overview.
For bills in non-supported languages, only the image is normally transmitted. The missing data is then filled in with the default data. Bills in partially supported languages can be processed by OCR, but not manually validated.
One currency is sourced from the bill. If there are several currencies on one bill, the home currency is selected. Similarly, only one bank account is read from the bill.
Bills received to Staria Capture service before 8.00 am on working days are processed no later than 12 pm on the next working day.