Context
An auto-parts manufacturing company in Córdoba, Argentina processed thousands of invoices each month across hundreds of suppliers. Matching invoices against purchase orders and goods receipts was fully manual.
The problem
Manual matching: 40 minutes per invoice
Analysts had to open invoices, find the PO, check the goods receipt, compare line items, and resolve differences by email.
- Identify the referenced PO
- Search the ERP
- Compare product, quantity, unit price, and taxes
- Validate received quantities
- Approve, reject, or request correction
Errors that cost money
- Duplicate payments when the same invoice arrived through different channels
- Overbilling when invoiced quantities exceeded received quantities
- Incorrect prices not detected before payment
- Credit notes not tied to the original invoice
Impact in numbers
- $15,000 USD/month in duplicate payments and overbilling
- 18-day average payment cycle
- 3 full-time people dedicated to matching
- 34% of invoices requiring investigation
The solution
Phase 1: Smart capture of the 3 documents
Invoices
Captured from email, portals, and electronic tax sources.
Purchase orders
Synchronized from the ERP with status, prices, and open balances.
Goods receipts
Captured from the ERP or scanned by warehouse teams.
Phase 2: Intelligent automated matching
The 3-way match
The system compares invoice ↔ PO, invoice ↔ goods receipt, and PO ↔ goods receipt.
Configurable tolerance rules
Rules define accepted differences for price, quantity, partial deliveries, and FX.
Workflows by result
Perfect matches auto-approve; partial matches route to analysts; missing documents trigger requests.
Phase 3: Exception management
For exceptions, Cedalio shows the three documents side by side, highlights differences, suggests the likely cause, and records the resolution.
Results
97% automatic match
Most invoices no longer require manual review.
From 40 minutes to 30 seconds
Processing time dropped dramatically for matched invoices.
Zero duplicate payments
Duplicate detection became part of the workflow.
Payment cycle: from 18 to 3 days
Approved invoices moved to payment faster.
Total savings
The company reduced leakage, manual effort, and supplier follow-up.
Technical details
How hard cases are resolved
Partial deliveries
The system supports partial invoices against open POs.
Suppliers that do not reference the PO
AI agents infer likely matches from supplier, items, dates, and amounts.
Credit notes
Credit notes are linked to the original invoice and exception history.
Service invoices without goods receipt
Alternative approval workflows handle services where no physical receipt exists.
Lessons learned
1. Manual 3-way matching does not scale
2. Tolerances are critical
3. The warehouse is part of the process
4. Quick wins build trust
Does your AP team spend hours matching invoices against POs and receipts?
Book a demo to see automated 3-way matching in action.