PO Matching in Accounts Payable

PO matching encompasses the automated validation of purchase order-backed invoices through systematic comparison of invoice data against authorized purchase orders and delivery receipts. The process includes purchase order detection, line-level matching algorithms, tolerance-based variance analysis, and exception routing for discrepancies that require human judgment.

The scope extends from basic 2-way matching between purchase orders and invoices to comprehensive 3-way matching that incorporates receipt verification, ensuring that payments align with both purchasing authorization and actual delivery confirmation.

Purchase order detection automatically identifies PO numbers on incoming invoices and retrieves corresponding purchase order data from the ERP system. The system scans invoice content for PO references, validates detected numbers against active purchase orders, and establishes the foundation for line-level matching processes.

This detection process handles various PO number formats and locations on invoices while maintaining accuracy through validation against the ERP's purchase order database. When multiple POs appear on a single invoice, the system manages partial matching scenarios and routes complex cases appropriately.

Two-way matching compares invoice line items directly against purchase order lines, validating quantities, unit prices, descriptions, and extended amounts. The system evaluates each invoice line against corresponding PO lines, applying matching logic that accounts for description variations, quantity tolerances, and price thresholds.

The process identifies exact matches, acceptable variances within configured tolerances, and exceptions requiring review. Invoice lines that match within established parameters proceed automatically, while discrepancies route to designated reviewers with clear variance explanations and supporting documentation.

Three-way matching adds receipt verification to the purchase order and invoice comparison, ensuring that payments align with actual deliveries. The system compares invoice quantities against both authorized PO quantities and confirmed receipt quantities, identifying discrepancies in any of the three documents.

This comprehensive validation prevents payment for undelivered goods and services while supporting complex scenarios such as partial deliveries, over-shipments, and multiple receipts against single purchase orders. Receipt data integration maintains real-time accuracy and supports audit trail requirements.

Tolerance thresholds define acceptable variances between purchase orders, receipts, and invoices without requiring manual intervention. Organizations configure percentage and dollar amount tolerances for quantities, unit prices, and extended amounts based on their risk tolerance and operational requirements.

The system applies these thresholds consistently across all matching scenarios, automatically approving invoices within tolerance while routing exceptions for review. Variance management includes escalation rules, approval requirements, and documentation standards for out-of-tolerance situations.

Exception handling manages invoices that fail automated matching criteria, providing clear variance explanations and resolution workflows. The system categorizes exceptions by type, such as quantity variances, price discrepancies, missing receipts, or unmatched line items, enabling targeted resolution approaches.

Resolution workflows support various outcomes including invoice correction, purchase order amendments, receipt adjustments, and manual approval overrides. Each resolution maintains audit trail documentation and feeds back into matching algorithm improvements.

Line-level matching algorithms compare individual invoice lines against purchase order lines using multiple matching criteria including item descriptions, quantities, unit prices, and account codes. The system handles description variations through semantic matching while maintaining accuracy in numerical comparisons.

Advanced matching logic accommodates common scenarios such as consolidated billing, partial shipments, and vendor-specific formatting differences. The algorithms continuously improve through machine learning while maintaining transparent matching decisions for audit purposes.