Hello @Wendy_2021 ,
I am confused now cause you mentioned PAD (which is Power Automate Desktop - the RPA solution of MS).
Your task is solvable with a usual Power Automate Flow.
I assume the document number is on the same position in every pdf document?
You can use the AI Builder and teach the model to extract exactly the information you need.

Have you already started to create your flow? If yes you could show me some screenshots so I can support you better.
With regards
Kevin