Hey @Long_Vu_Nguyen, thanks so much for the feedback!
I don’t know if you’ve seen my latest post yet, but I’ve already released a second version of the PO workflow. It now includes improved validation for fields that couldn’t be extracted correctly, duplicate PO number detection to prevent the same document from being processed twice, and I’m currently collaborating with another builder who created a tool for logging extracted values before they reach downstream business systems. That makes it much easier to trace where every piece of data originated if something ever needs to be audited or debugged.
If you’re interested, feel free to check it out here: