I’m aware of some limitations within the AI Builder, but I’m hoping someone here can help me work around a particular challenge I’m facing.
I’m working with multipage PDFs, some of which are over 200 pages long. Each page contains one of eight different types of documents (e.g., time worked data, payroll data, etc.). While the documents differ slightly, I can extract the necessary information as tables using the AI Builder. However, I’m encountering an issue with associating two critical pieces of data with these tables:
- Company Name: This appears at the top center of the page and varies from page to page. A single PDF may contain documents from multiple companies.
- Date ("For the pay period ending"): This appears once before the table on the relevant pages.
I understand that one of the AI Builder’s limitations is that it only allows tagging data once per table. This means I can't tag data for each row of the table or tag the company name and date fields multiple times across the pages.
I’m considering using a separate AI model to extract the Page/Company Name/Date information and then running it in parallel with the main AI model that extracts the table data. However, I’m not sure how to set this up effectively.
Has anyone encountered a similar challenge or have any suggestions on how to approach this? Any guidance on running two models in parallel or other workarounds would be greatly appreciated!
Thank you in advance for your help!
