Hi! I'm working with Laserfiche Workflow 12 locally. I receive a very large PDF (sometimes over 1,000 pages) with invoices from multiple vendors and occasionally messy pages (I can't use fixed areas or templates).
What I've set up so far:

• Retrieve the text from the document.
• C# script that parses the text and generates lines like: INV=123456|PAGES=2-5. If a page doesn't have a SAP, I deduce the group by matching (PO, customer, or vendor + total) and inherit the SAP for the rest of the group.
• Pattern matching + For Each to iterate each line and copy/move pages:
• If there is a SAP → I create/use \Invoices\{SAP} and compile the final document.
• If there are no clear references → I send it to \Invoices\Manual Review.

It works, but it's complex to maintain, and I want to simplify it, reduce the code, and make it design-agnostic. Do you recommend any recipes or patterns for the local platform to sort by page and then rebuild with Workflow, or a script-free flow to:

• extract keys by page (SAP/PO/Customer/Vendor/Total) without fixed zones,
• propagate keys detected on one page to the rest of the group,
• handle ambiguities (repeated POs, "Total" on multiple pages),
• and decide whether to copy or move pages in large batches?

If anyone has a checklist, example, or QF session/Workflow reference, I'd greatly appreciate it! 🙌