How to add document OCR to Macola
Short answer
Run OCR and field extraction on invoices, packing slips and certificates, match the results to open POs and receipts in Macola's database, and route low-confidence fields to a reviewer before anything posts. Keep the documents and models in-house when the content is sensitive.
Macola's accounts payable and receiving teams still handle a large amount of paper and PDF: vendor invoices, packing slips, freight bills, mill certs and certificates of conformance. Each one needs to be matched to a purchase order, keyed and filed. The work is repetitive, but errors are costly because a wrong invoice amount or quantity creates payment disputes.
OCR on its own only turns pixels into text. The value comes from the layer that knows Macola: that a PO number maps to a poordhdr-style record, that a vendor has specific remit-to data and that a three-way match depends on received quantity.
The design that works is conservative. Extract, validate against the database, show a reviewer the document next to the fields, and post only what passes.
What teams use it for
Vendor invoice capture
Extract vendor, invoice number, date, totals and PO reference, then match to open PO lines and propose an AP voucher for review.
Packing slip to receipt
Read packing slips at the dock, match item and quantity to the PO, and prepare a receiving entry for confirmation.
Certificate of conformance filing
Extract lot or heat numbers and PO references from certs and attach them to the receipt record or document store.
Freight bill matching
Match carrier invoices to shipments and flag charges outside expected ranges.
Customer PO scanning
Convert scanned customer POs into draft order lines with item cross-reference lookups.
Drawing and revision capture
Index drawing numbers and revisions from attachments so engineers can find the right document for an item.
Macola integration surface
- Open PO and receipt data in SQL Server provides the reference for three-way matching; confirm exact tables for your version.
- Vendor master data supports validation of remit-to addresses and payment terms.
- AP entry and receiving are the target transactions. Use a licensed web service, import routine or supported entry screen automation; avoid direct table inserts.
- Macola document management or attachment features vary by version and add-ons; if absent, store files in a network share keyed by document number.
- An integration-owned database holds OCR text, field confidences, reviewer actions and file hashes.
How it works
- 1
Ingest documents
Pull from a scan folder, scanner-to-email address or a mailbox, and compute a file hash to prevent double processing.
- 2
Run OCR and layout analysis
Use a local OCR engine for scans and direct text extraction for digital PDFs, keeping bounding boxes for review.
- 3
Extract structured fields
Pull header and line items with rules per vendor where formats are stable and a model for variable layouts.
- 4
Match to Macola
Look up the PO, vendor and received quantities, and check price, quantity and tax tolerances.
- 5
Score and route
High-confidence, fully matched documents go to quick approval. Anything else lands in an exception queue.
- 6
Review interface
Show the image beside the fields with highlighted source regions so a clerk can correct in one place.
- 7
Post and archive
Create the voucher or receipt through the supported interface and link the original file to the transaction.
Example flows
Invoice three-way match
Trigger: A vendor invoice PDF arrives in the AP mailbox.
- The pipeline extracts invoice number, PO number, lines and total.
- It checks the PO and receipt quantities in Macola.
- Price and quantity fall within tolerance, so the voucher is proposed.
- An AP clerk approves and the voucher is created.
Result: Clean invoices are processed with a quick check, and the clerk's time goes to exceptions.
Dock receipt with cert
Trigger: A packing slip and certificate are scanned at receiving.
- OCR reads the slip and matches the PO lines.
- Heat or lot numbers are extracted from the cert.
- The receiver confirms quantities on a tablet and the receipt is staged.
Result: Receiving records carry traceability data without separate manual entry.
Keeping data on-prem
Invoices, certs and drawings frequently contain pricing, customer and technical data that you may not want in an external OCR cloud. Run OCR and extraction on a server in your own network using open-source OCR engines and an open-weight vision or language model, and keep the document store on internal storage. The Macola connection needs read access to PO, receipt and vendor data and a controlled write path for approved vouchers and receipts. For air-gapped or restricted sites, package models and OCR dependencies for offline installation and update them through a controlled process.
Security checklist
- Store documents in encrypted internal storage with role-based access and retention rules.
- Validate file types, scan for malware and treat document text as untrusted data, not instructions.
- Use a read-only SQL role for matching and a single approved write path for posting.
- Keep bank detail and remit-to changes behind a separate human verification step to prevent payment fraud.
- Log reviewer corrections with user, time and before/after values.
Ways to get it done
Off-the-shelf AP automation or capture product
Effort: Weeks, plus subscription and mapping work.
Fits shops that want a packaged AP workflow and have a vendor with a Macola connector or file-based import.
Build in-house with open-source OCR
Effort: Roughly four to eight weeks for invoices with review UI.
Fits teams with development resources and a few high-volume vendors with consistent formats.
Partner build with Netray
Effort: A pilot on one document type typically takes a few weeks.
Fits mixed document types, traceability needs and a requirement to keep documents on-prem.
Pitfalls
- !Trusting OCR totals without checking that lines sum to the total.
- !Posting vouchers without confirming received quantity, which defeats three-way match.
- !Treating vendor remit-to or bank changes found in a PDF as valid updates.
- !No duplicate-invoice detection by vendor, number and amount.
- !Training only on clean samples and failing on skewed or faxed scans.
Frequently asked questions
Does Macola include OCR?
Do not count on it as a core feature. Some versions or add-ons may offer document capture, so ask your reseller. Otherwise you add a separate OCR and extraction layer that feeds Macola through a supported interface.
How accurate is OCR on invoices?
It depends on scan quality and vendor layout. Digital PDFs are far better than scans. Measure field-level accuracy on your own documents, set confidence thresholds, and keep human review for exceptions.
Can this run without sending documents to the cloud?
Yes. Open-source OCR and local models run on your own servers, and the Macola connection stays inside your network. This suits sensitive pricing and defense-related documents.
How is three-way matching handled?
The pipeline compares invoice lines to PO lines and received quantities in Macola, applies price and quantity tolerances, and flags mismatches for a clerk instead of posting automatically.
Can it attach documents to Macola records?
Depending on your version and add-ons, attachments may be supported natively. If not, store files on a managed share named by document number and keep the link in your integration database.
Want Document OCR working on your Macola?
A free 30 minute assessment: we map the integration path for your version and setup, and where it should run.
Get a free assessment