Cash application is killing us right now and I’m hoping someone here has figured this out. We get remittance advice from customers in every format imaginable—email attachments, check stubs, portal exports, sometimes literally a photo someone took with their phone. Matching payments to open invoices manually is a nightmare and our AR team is buried.
Can OCR actually help here, or is the document variety too much of a problem? Curious whether anyone’s automated this and what the setup actually looked like.
Been there. Cash application was the bane of our AR team’s existence until we actually dug into automating it—and honestly it’s one of the highest-ROI things you can do with OCR if you get it right.
The core challenge is exactly what you described: the document chaos. Structured PDFs, loose scans, portal exports, phone photos of check stubs—you need something that doesn’t require you to build a separate workflow for every customer’s remittance format. That’s where template-based tools fall short. We tried one early on and spent more time maintaining templates than we saved.
What actually worked for us was switching to an AI-based extraction approach. We use Lido now and it handles the variety without needing predefined templates—pulls invoice numbers, customer reference codes, payment amounts, dates, all of it regardless of how the document is structured. That data flows into our accounting software and gets matched against open invoices automatically.
FWIW, here’s what I’d look for in any solution you evaluate: handles all input formats (PDF, images, email attachments, faxes), flexible field extraction since customers all format remittance differently, solid ERP/accounting integration, and good exception handling for the extractions it’s not confident about.
Results-wise—our DSO came down about 3 days, the team saves probably 30+ hours a month on manual entry, and cash application processing is dramatically faster. Setup took us around 2-3 weeks including validation against our top customers’ actual documents, which I’d recommend doing before you go live.
+1 on this. Template-based was a disaster for us too. Every time we onboarded a new customer with a slightly different remittance format it was like starting from scratch. Switched probably 18 months ago now and the maintenance burden is just… gone. Genuinely kind of hard to believe we put up with the old way for as long as we did.
Oh interesting, we were actually in a similar eval process earlier this year. Rossum was on our shortlist too. Can I ask — how did the spreadsheet integration actually work in practice? Like is it a native connector or did you have to build something custom? We’re also a heavy Google Sheets shop so that caught my eye.
We ended up going a different direction mostly because of pricing but honestly second-guessing it a bit reading this thread. Would love to hear how the pilot held up once you went into full production.
Jumping in here because we literally just went through this with Lido last month. Multi-page stuff works fine in our experience — it stitches the pages together and treats the whole thing as one document. That said, we did notice it occasionally missed line items that were split across a page break, so worth double-checking those edge cases if your vendors do running totals across pages. Overall though, way less painful than what we were doing before.