I am trying to create a custom AI model to extract information from a series of documents (multiple pages per document).
The documents are structured in a table form with a couple of complications.
First: The employee name for each table only appears once per table. When this information is extracted and saved to sharepoint, I need the employee name to appear on each row of the table instead of only one in order to identify the table. Is this possible?
Second: Though individual tables are not more than a page in length, some can start at the bottom of a page and run onto the next. Does this mean it is necessary to use the multi-page function or can the tables be accurately extracted normally with the right amount of training?
I have attached an example document below. Any help is greatly appreciated.
Hi @CSilva ,
First point: If you are using a Power Automate flow, that would not be an issue to reuse the same field for all rows added in your SharePoint. Can you do a screenshot of your flow, and particularly the part where you insert data in your SharePoint?
Second point: Depending on the format of your documents, yes, you would need to use the multi-page table option. Look at this documentation that explains how data can be extracted when there are multipage tables: Process multipage tables in document processing in AI Builder (experimental) | Microsoft Learn
Let us know if you have more questions.