AI Order Data Reconciliation Agent for Industrial Fittings and Components

How Cassidy automates EDI three-way matching using AI

Step 1: Ingest EDI documents and normalize data

The Workflow triggers when new 850 (PO), 856 (ASN), or 810 (Invoice) documents arrive via AS2, VAN, or SFTP. Cassidy normalizes trading partner mappings, cross-references customer part numbers to internal SKUs, and validates UOM, pack sizes, and pricing basis against your Knowledge Base.

Step 2: Execute three-way match with tolerance rules

Cassidy compares PO terms against shipped quantities from the ASN and invoiced amounts, applying your configured tolerances for price variance, quantity variance, and allowable charges. Matches within tolerance are auto-approved; outliers are flagged with specific reason codes.

Step 3: Validate traceability and compliance artifacts

For industrial fittings requiring MTRs, Certificates of Conformance, or heat/lot numbers, Cassidy verifies these documents are present and linked to the correct line items, preventing receipt blocks and downstream audit issues.

Step 4: Route exceptions to the right team

Mismatches are categorized by type—price discrepancy, missing ASN, SSCC error, or QA non-conformance—and routed to the appropriate owner (AP, customer service, logistics, or quality) with full context: annotated document comparisons, carton contents, and supporting evidence.

Step 5: Human-in-the-loop disposition

Your team reviews flagged exceptions in a side-by-side view with one-click options to accept, short pay, rebill, or escalate. Cassidy captures the decision and updates downstream systems accordingly.

Step 6: Post to ERP and update scorecards

Clean matches and resolved exceptions flow directly to your ERP for vouchering and GL posting. Cassidy updates variance accounts and feeds compliance dashboards with metrics like auto-match rate, ASN timeliness, PPV trends, and partner scorecards.