Hi,
I would like to create a document processing process, but on the same service email address I receive multiple document types. Because of the language (Romanian) and some specific fields I cannot use the prebuilt models, and I have to create custom models for each document type.
Let's say there are three model types:
- Invoices
- Receipts
- Activity Reports
From what I've found until now - one approach is to use the "predict" action with each of the model. The problem with this approach is that will consume a lot of AI Builder credits, especially when the model types will increase.
I cannot detect document type based on sender, because the same sender could attach different document types, so the detection should be at the document level.
So, is there a way to determine the "document type" and after that to use the right model for it? Which AI Builder model would better achieving this task: Entity Extraction, Category classification or else (assuming that I probably need to use "Text recognition" along the way)?
Kind Regards,
Lucian