Manufacturing AP Exception Handling: Partial Shipments, Substitute Materials & Quality Rejections (2026)

---

TL;DR

---

Why Manufacturing AP Exceptions Are Different (And Harder)

Manufacturing vs. Service Industry Exception Rates

Service Company Exception Profile:

• Total invoices monthly: 1,000
• Exception rate: 25-30%
• Exception types: Price variances (contract amendments), missing PO numbers, approval routing errors
• Resolution complexity: Moderate (email procurement, verify contract, approve)

Manufacturing Company Exception Profile:

• Total invoices monthly: 1,200
• Exception rate: 45-55%
• Exception types: Partial shipments, substitute materials, quality rejections, batch variations, over-delivery, damaged goods, tool/die charges, packaging variances
• Resolution complexity: High (requires warehouse confirmation, engineering validation, quality inspection reports, vendor credits reconciliation)

The Difference: Manufacturing exceptions require operational context beyond financial data.

Service industry exception:

“Vendor increased price from $100 to $110. Check contract amendment. Approve if legitimate.”

Manufacturing exception:

“Vendor shipped Part# ABC-200 instead of ABC-100. Invoice shows 5% price increase. Is ABC-200 an approved substitute? Does engineering change order exist? Does material spec match? Is price increase justified for higher-grade material? Did quality inspection pass? Should PO be updated for future orders?”

Manufacturing exceptions require cross-functional validation (engineering, quality, warehouse, procurement) that traditional AP automation cannot orchestrate.

---

Manufacturing Exception Taxonomy: The Seven Common Scenarios

Exception Type 1: Partial Shipments Due to Production Capacity

What Happens:

PO: 500 units Component XYZ at $45/unit = $22,500

Vendor reality:

• Production line capacity: 300 units/week
• Lead time for 500 units: 2 weeks
• Customer needs material ASAP for production schedule

Vendor solution:

• Week 1: Ship 300 units immediately (partial shipment #1)
• Week 2: Ship remaining 200 units (partial shipment #2)

Invoicing:

• Invoice #1: 300 units × $45 = $13,500
• Invoice #2 (1 week later): 200 units × $45 = $9,000

Three-Way Matching Result:

Field	PO	GR (Week 1)	Invoice #1	Match?
Quantity	500	300	300	- MISMATCH (invoice 60% of PO)
Amount	$22,500	N/A	$13,500	- MISMATCH
Result				- EXCEPTION FLAGGED

Manual Investigation Required:

1. AP emails warehouse: “Did we receive 300 units or 500 units?”
2. Warehouse responds (1-2 days): “Received 300 units. Delivery note mentions ‘Partial shipment 1 of 2, backorder 200 units’”
3. AP emails procurement: “Was partial shipment authorized?”
4. Procurement responds (1 day): “Yes, vendor requested split shipment due to capacity. Approved.”
5. AP creates variance documentation and approves $13,500 invoice
6. AP updates PO: 300 units invoiced, 200 units open

Time: 3-5 days

AI Resolution:

• AI detects 40% quantity variance
• Checks goods receipt: 300 units received -
• Finds Advanced Shipping Notice (ASN) in vendor email: “Partial shipment 1 of 2”
• Confirms PO allows partial shipments
• Auto-approves invoice for 300 units
• Updates PO status: 300 invoiced, 200 backorder
• Sets flag: Expect invoice #2 when backorder ships

Time: 8 minutes

Exception Type 2: Substitute Materials (Approved Alternates)

What Happens:

PO: 1,000 units Steel Sheet ASTM A36, 4mm thickness, Part# STEEL-A36-4MM at $120/unit

Vendor Situation:

• ASTM A36 steel mill experiencing production delays (raw material shortage)
• Vendor has ASTM A572 (higher strength grade) in stock
• A572 costs more ($125/unit) but meets customer’s engineering specs

Vendor Action:

• Contacts customer engineering: “Can we substitute A572 for A36? Meets your load requirements.”
• Engineering approves: “Yes, A572 acceptable. Update drawings to reflect material change.”
• Engineering creates ECO (Engineering Change Order) #2026-0095: “Part# STEEL-A36-4MM superseded by STEEL-A572-4MM”

Invoicing:

• Vendor ships: 1,000 units ASTM A572, Part# STEEL-A572-4MM
• Vendor invoices: Part# STEEL-A572-4MM, 1,000 units × $125 = $125,000

Three-Way Matching Result:

Field	PO	Invoice	Match?
Part Number	STEEL-A36-4MM	STEEL-A572-4MM	- MISMATCH
Unit Price	$120	$125	- MISMATCH (4.2% increase)
Total	$120,000	$125,000	- MISMATCH
Result			- EXCEPTION FLAGGED

Manual Investigation:

1. AP sees part number mismatch and price variance
2. AP emails procurement: “Vendor shipped wrong part and increased price. Error or authorized?”
3. Procurement emails engineering: “Did you approve A572 substitute?”
4. Engineering responds (2-3 days): “Yes, ECO-2026-0095 approved substitution. Update PO.”
5. Procurement forwards ECO to AP
6. AP reviews ECO, confirms substitution legitimate
7. AP checks if 4.2% price increase reasonable for higher-grade material
8. AP creates variance approval request
9. Approvers review and approve (1-2 days)
10. AP manually releases invoice

Time: 5-7 days

AI Resolution:

• AI detects part number mismatch + 4.2% price variance
• Searches ECO database for “STEEL-A36-4MM”
• Finds ECO-2026-0095: “Approved substitute: A572 for A36, effective immediately”
• Validates material specs compatible
• Confirms price increase <5% (within tolerance for material upgrade)
• Checks goods receipt: 1,000 units A572 received and accepted by quality
• Auto-approves invoice with documentation: “Material substitution per ECO-2026-0095”
• Notifies procurement: “Update PO part number to A572 for future orders”

Time: 15 minutes

Exception Type 3: Quality Rejections with Vendor Credits

What Happens:

PO: 5,000 units Machined Component ABC at $25/unit = $125,000

Receiving Process:

• Vendor ships 5,000 units
• Warehouse receives shipment, creates goods receipt for 5,000 units
• Quality team performs inspection sampling (per ISO 9001 procedures)
• Inspection finds: 4,650 units pass spec, 350 units fail (dimensional tolerance out of range)

Vendor Communication:

• Quality team notifies vendor: “350 units rejected due to dimensional defects”
• Vendor requests return of defective units for rework
• Vendor issues credit memo: 350 units × $25 = $8,750

Invoicing:

• Original invoice: 5,000 units × $25 = $125,000
• Vendor credit memo: -$8,750
• Net invoice amount: $116,250 (4,650 accepted units)

Three-Way Matching Result:

Field	PO	GR (Initial)	Net Invoice	Match?
Quantity	5,000	5,000	4,650	- MISMATCH (7% under)
Amount	$125,000	N/A	$116,250	- MISMATCH
Result				- EXCEPTION FLAGGED

Manual Investigation:

1. AP sees invoice for 4,650 units but goods receipt shows 5,000 received
2. AP emails warehouse: “Why is invoice 350 units less than goods receipt?”
3. Warehouse responds: “Check with quality team. We received 5,000.”
4. AP emails quality: “Did you reject any units from this shipment?”
5. Quality responds (1-2 days): “Yes, 350 units failed inspection. See report QC-2026-0234.”
6. AP requests quality inspection report
7. AP verifies vendor credit memo received for rejected quantity
8. AP matches: 350 rejected units = 350 unit credit = invoice reduced by 350 units -
9. AP approves net invoice $116,250

Time: 3-5 days

AI Resolution:

• AI detects 7% quantity variance (invoice 4,650 vs. GR 5,000)
• Searches quality management system for inspection report on this shipment
• Finds QC-2026-0234: “Inspected lot ABC-2026, 350 units failed dimensional tolerance”
• Pulls vendor credit memo: -$8,750 for 350 rejected units
• Validates: 5,000 received - 350 rejected = 4,650 accepted = invoice quantity -
• Confirms credit amount: 350 × $25 = $8,750 -
• Auto-approves net invoice $116,250
• Documents: “Invoice adjusted for quality rejection per QC-2026-0234, vendor credit applied”

Time: 10 minutes

Exception Type 4: Batch Variation Tolerances

What Happens:

PO: 10,000 liters Industrial Chemical XYZ at $12/liter = $120,000

Vendor Delivery:

• Vendor fills tanker truck from storage tank
• Tank measurement accuracy: ±1.5% (industry standard for liquid bulk materials)
• Actual delivered: 10,150 liters (1.5% over PO quantity)
• Vendor invoices actual delivered quantity

Invoicing:

• Invoice: 10,150 liters × $12 = $121,800 (1.5% over PO amount)

Three-Way Matching Result:

Field	PO	GR	Invoice	Match?
Quantity	10,000 L	10,150 L	10,150 L	- MISMATCH (1.5% over)
Amount	$120,000	N/A	$121,800	- MISMATCH
Result				- EXCEPTION FLAGGED

Traditional ERP Logic:

• Exact match required: 10,150 ≠ 10,000 → EXCEPTION
• Even though 1.5% variance is industry-standard tolerance for bulk liquids

Manual Investigation:

1. AP sees 1.5% over-delivery and price variance
2. AP emails warehouse: “Did we receive 10,150 liters or 10,000 liters?”
3. Warehouse responds: “Tanker delivered 10,150 L per meter reading. Normal variance for bulk liquid.”
4. AP emails procurement: “Is 1.5% over-delivery acceptable?”
5. Procurement responds: “Yes, industry standard ±2% tolerance for bulk chemicals.”
6. AP approves invoice for actual quantity delivered

Time: 2-3 days

AI Resolution:

• AI detects 1.5% quantity over-delivery
• Identifies material category: Bulk liquid chemical
• Applies industry-standard tolerance: ±2% for bulk liquids
• Calculates actual variance: (10,150 - 10,000) / 10,000 = 1.5% - WITHIN TOLERANCE
• Confirms goods receipt quantity matches invoice: 10,150 L -
• Auto-approves invoice for actual delivered quantity
• Documents: “Batch variation 1.5% within ±2% tolerance for bulk liquid materials”

Time: 5 minutes

Exception Type 5: Packaging Minimum Over-Delivery

What Happens:

PO: 480 units Component DEF at $15/unit = $7,200

Vendor Packaging:

• Component DEF ships in boxes of 100 units
• 480 units would require 4 full boxes (400 units) + partial box (80 units)
• Vendor policy: No partial boxes (shipping/handling efficiency)

Vendor Action:

• Ships 5 full boxes = 500 units
• Invoices for actual shipped: 500 units × $15 = $7,500

Three-Way Matching Result:

Field	PO	GR	Invoice	Match?
Quantity	480	500	500	- MISMATCH (4.2% over)
Amount	$7,200	N/A	$7,500	- MISMATCH

Manual Investigation:

1. AP sees 20 unit over-delivery (4.2%)
2. AP emails warehouse: “Why did vendor ship 500 instead of 480?”
3. Warehouse: “Vendor ships in boxes of 100. Received 5 boxes = 500 units.”
4. AP emails procurement: “Should we pay for 500 or 480?”
5. Procurement: “Packaging minimum is standard for this vendor. Approve 500 units. We can use extras.”
6. AP approves

Time: 2-3 days

AI Resolution:

• AI detects 4.2% over-delivery (20 units)
• Checks vendor master data: “Packaging = 100 units/box”
• Calculates: 480 units = 4.8 boxes → rounds to 5 boxes = 500 units
• Validates: Over-delivery matches packaging minimum -
• Confirms over-delivery <5% (within typical tolerance) -
• Checks if company accepts over-delivery for this commodity (historical pattern: yes)
• Auto-approves invoice for 500 units
• Documents: “Over-delivery due to vendor packaging minimum (100 units/box)”

Time: 6 minutes

---

How AI Resolves Manufacturing Exceptions: Required Integrations

Integration 1: Warehouse Management System (WMS)

Critical Data from WMS:

• Goods receipt records (actual quantities received)
• Receiving inspection notes (“Partial shipment 1 of 2”, “Damaged units: 15”)
• Delivery documentation (packing slips, bills of lading, delivery notes)
• Warehouse acceptance/rejection status
• Physical inventory confirmation

AI Use Cases:

• Match invoice quantity to actual received quantity (not PO quantity)
• Validate partial shipments against delivery notes
• Confirm damaged goods documented before approving vendor credits

Integration 2: Quality Management System (QMS)

Critical Data from QMS:

• Inspection reports (pass/fail results, defect types, quantities)
• Vendor quality ratings (defect rates, on-time delivery, compliance)
• Material certifications required (mill certs, COC, material test reports)
• Quality holds (shipments awaiting inspection, quarantined lots)
• Non-conformance reports (NCR) documenting quality issues

AI Use Cases:

• Validate quality rejections before approving reduced invoice quantities
• Verify vendor provided required material certifications
• Check if vendor quality performance warrants stricter invoice scrutiny

Integration 3: Engineering Change Management (ECM)

Critical Data from ECM:

• Engineering Change Orders (ECO) documenting approved modifications
• Approved substitute materials (part A can replace part B)
• Bill of Materials (BOM) alternates
• Specification updates (material grades, dimensions, tolerances)

AI Use Cases:

• Validate substitute materials approved by engineering before approving price variances
• Check if part number mismatches are legitimate ECO-driven changes
• Confirm material spec changes authorized

Integration 4: Production Planning System

Critical Data:

• Expected delivery schedules
• Partial shipment authorizations
• Backorder status and expected ship dates
• Production urgency (JIT requirements, line-down situations)

AI Use Cases:

• Validate partial shipments against authorized delivery schedules
• Prioritize invoice processing for critical materials needed for production
• Set expectations for backorder invoice arrival

---

Real-World Manufacturing Exception Resolution: Before and After AI

Company Profile

• Industry: Automotive component manufacturing
• Monthly invoice volume: 1,200 invoices
• Exception rate: 52% (625 exceptions monthly)
• AP team size: 3 FTE
• Exception types: 35% partial shipments, 25% substitute materials, 18% quality rejections, 12% batch variations, 10% other

Before AI (Manual Exception Resolution)

Time Allocation:

• Exception investigation: 32 hours/week across team
• Email follow-up (warehouse, quality, engineering, procurement): 18 hours/week
• Variance approval documentation: 8 hours/week
• Total exception effort: 58 hours/week = 1.45 FTE dedicated to exceptions

Financial Impact:

• Late payment due to exception delays: 18-22 days average
• Lost early payment discounts: $195,000 annually (missed 2% 10 net 30 on $10M material spend)
• Vendor escalation calls: 25 monthly
• Payment terms at risk: 3 key suppliers threatened payment-on-receipt due to chronic delays

Operational Pain Points:

• Production delays: 2 instances where vendor delayed shipment due to unpaid invoices
• AP team burnout: High turnover, difficulty filling positions
• Month-end close: 3 extra days due to exception backlog accruals

After AI Implementation (8 Months Post-Deployment)

Automation Results:

Exception Type	Monthly Volume	AI Auto-Resolution Rate	Manual Review Required
Partial shipments (with ASN)	220	92% (202 invoices)	18 invoices
Substitute materials (approved ECO)	155	88% (136 invoices)	19 invoices
Quality rejections (with QC report)	115	94% (108 invoices)	7 invoices
Batch variations (within tolerance)	75	97% (73 invoices)	2 invoices
Packaging minimums	35	91% (32 invoices)	3 invoices
Other exceptions	25	65% (16 invoices)	9 invoices
TOTAL	625	90% (567 invoices)	58 invoices

Time Savings:

• Exception investigation: 32 hours/week → 6 hours/week (81% reduction)
• Email follow-up: 18 hours/week → 2 hours/week (89% reduction)
• Manual variance documentation: 8 hours/week → 1 hour/week (88% reduction)
• Total exception effort: 9 hours/week (vs. 58 hours) = 49 hours/week saved

Financial Impact:

• Early payment discount capture: $195K lost → $28K lost = $167K annual recovery
• Late fees eliminated: $8,500 annual savings
• Vendor escalation calls: 25/month → 4/month
• Payment terms: Restored preferred supplier status with all vendors

Operational Improvements:

• Production continuity: Zero instances of shipment delays due to payment issues
• AP team morale: Reduced burnout, staff focusing on vendor relationships and process improvement
• Month-end close: Reduced by 2 days (exception backlog eliminated)

ROI Calculation:

Annual Benefits:

• Labor savings: 49 hours/week × 50 weeks × $35/hour = $85,750
• Early payment discount recovery: $167,000
• Late fee elimination: $8,500
• Total annual benefit: $261,250

Implementation Cost:

• Year 1: $45,000 (software + implementation + training)
• Year 2+: $28,000 (annual licensing)

Year 1 ROI: 481% | Payback: 2.1 months

---

Our Verdict: Manufacturing Requires AI-Powered Exception Handling

Key Insights

1. Manufacturing Exception Rates Are 50-80% Higher

Service companies: 25-35% exception rate Manufacturing companies: 45-55% exception rate

Reason: Operational complexity (partial shipments, substitutions, quality issues) unavoidable in manufacturing supply chains.

2. Traditional ERP Cannot Handle Operational Exceptions

ERP three-way matching is binary: exact match or exception.

Manufacturing reality: Legitimate variances (batch tolerances, packaging minimums, quality rejections) flagged as exceptions despite being operationally valid.

3. AI Requires Manufacturing System Integration

Generic AP automation won’t suffice. AI must connect to:

• WMS (goods receipt context)
• QMS (quality inspection results)
• ECM (engineering approvals for substitutes)
• Production planning (partial shipment authorizations)

4. Exception Resolution Speed Directly Impacts Vendor Relationships

Manufacturing depends on reliable supplier partnerships. Chronic late payments due to exception delays damage:

• Payment terms (vendors demand shorter payment windows)
• Supply priority (vendors allocate scarce materials to faster-paying customers first)
• Pricing leverage (vendors less willing to negotiate discounts with chronically late payers)

AI exception resolution preserves supplier relationships by enabling on-time payment despite high exception rates.

---

How Peakflo Handles Manufacturing AP Exceptions

Peakflo’s agentic AI platform includes manufacturing-specific exception resolution agents that integrate with WMS, QMS, and ECM systems to autonomously validate operational variances.

Peakflo Manufacturing Capabilities

Manufacturing System Integrations:

• WMS: SAP EWM, Oracle WMS, Manhattan, Blue Yonder, NetSuite WMS
• QMS: ETQ Reliance, Sparta Systems TrackWise, MasterControl, Qualio
• ECM: Windchill, Teamcenter, Arena PLM, Propel

Exception Resolution Agents:

• Partial shipment agent: Validates ASN, backorder schedules, goods receipts
• Substitute material agent: Checks ECO database, validates material equivalency, confirms pricing reasonableness
• Quality rejection agent: Matches inspection reports to vendor credits, validates rejected quantities
• Batch variation agent: Applies industry-standard tolerances by material category
• Multi-line invoice agent: Processes mixed exceptions, enables partial approval

Confidence-Based Processing:

• High confidence (88%+): Auto-approve with manufacturing documentation
• Medium confidence (75-87%): Auto-approve with AP notification
• Low confidence (<75%): Escalate with operational analysis and recommendation

Auto-Resolution Rates:

• Partial shipments: 90-94%
• Substitute materials: 85-90%
• Quality rejections: 92-96%
• Batch variations: 96-98%
• Overall manufacturing exceptions: 75-85%

Peakflo Implementation for Manufacturing

Week 1-2: Manufacturing system discovery (WMS, QMS, ECM), exception type analysis Week 3-4: System integrations (ERP + manufacturing systems via APIs) Week 5-6: AI training on historical exceptions (ECO patterns, quality rejection workflows, vendor behavior) Week 7-8: Pilot testing with live manufacturing invoices Week 9-10: Production rollout, continuous optimization

Time to value: 8-10 weeks from kickoff to 75%+ exception auto-resolution

Related Peakflo Resources

• Three-Way Matching Exception Resolution Guide
• AI Exception Management for Accounts Payable
• Agentic Workflows vs Traditional AP Automation
• AI GL Coding Automation
• Accounts Payable Automation ROI

---

Frequently Asked Questions

What are manufacturing-specific AP exceptions?

Manufacturing AP exceptions are invoice variances unique to production environments: (1) Partial shipments - vendor ships 300 of 500 ordered units due to production capacity, backorders remaining units, (2) Substitute materials - vendor ships alternate part (higher grade steel, different batch, upgraded component) when original unavailable, (3) Quality rejections - goods received but fail quality inspection, vendor issues credit for defective units, (4) Batch variations - ingredient quantities vary by batch (±2-3% for liquids, powders), creating quantity variances, (5) Scrap allowances - vendor ships 5% extra units accounting for expected manufacturing scrap, (6) Over-delivery for packaging minimums - vendor ships 520 units instead of 500 to fill complete pallet. These scenarios occur in 45-55% of manufacturing invoices vs. 25-35% in service industries.

Why do partial shipments happen in manufacturing?

Partial shipments occur due to five manufacturing realities: (1) Production capacity constraints - vendor’s line can only produce 300 units/week, splits 500-unit order across two deliveries, (2) Material shortages - vendor receives insufficient raw materials, ships what’s available immediately, backorders remainder, (3) Quality holds - 200 of 500 units fail vendor’s internal QC, vendor ships 300 passing units, reworks failed batch, (4) Logistics optimization - vendor ships partial quantity to meet customer production deadline rather than delaying full shipment, (5) Long-lead items - standard components ship immediately, custom components with 8-week lead ship separately. For manufacturers with JIT (just-in-time) production, partial shipments create invoicing complexity: invoice for 300 units at $45 = $13,500 doesn’t match PO for 500 units at $45 = $22,500, requiring exception investigation.

How do substitute materials create AP exceptions?

Material substitutions create three-way matching failures: (1) Part number mismatch - PO specifies “Steel Grade A36”, vendor ships “Steel Grade A572” (higher strength), invoice shows different part number, (2) Price variance - substitute material costs more ($125/unit vs. $120 PO price) but vendor absorbs difference or charges premium, (3) Unit of measure changes - PO for “sheets”, vendor ships “coils” (different UOM), invoice quantity appears wrong. Real scenario: Electronics manufacturer orders capacitor Part# CAP-1000. Vendor discontinues CAP-1000, ships CAP-1200 (equivalent spec, different manufacturer). Invoice shows CAP-1200 at $0.85 vs. PO CAP-1000 at $0.80. AP clerk must verify: (1) Is CAP-1200 approved substitute? (Check engineering change order), (2) Is price variance acceptable? (5% increase justified?), (3) Should PO be updated for future orders? Manual investigation: 45-60 minutes per invoice.

---

Ready to automate manufacturing AP exception handling? Schedule a demo with Peakflo to see how AI resolves partial shipments, substitute materials, and quality rejections autonomously.