Invoice Extraction Failure Modes

Automated invoice extraction quality audit across document headers, line items, and reconciliation outcomes

Total Invoices
83
summary-sheet baseline
Total Line Items
161
flattened line records
Reconciliation Failures
17
20.5% of invoices
Missing Due Date
64
77.1% of invoices
KPI indicator strip
Totals and normalized failure rates across invoice headers and flattened line items.
Missing-field frequency by extraction step
Header-level fields are normalized to 83 invoices; line-item fields are normalized to 161 flattened lines.
Reconciliation outcomes
Share of invoices where subtotal plus tax matches the extracted grand total.
Line-item value buckets by reconciliation status
Histogram-style view using flattened line totals to show where failed records cluster by size.
Vendors with the highest issue concentration
Stacked counts combine missing quantity, missing unit price, and failed reconciliation labels in the flattened line view.
Key takeaways
Header
Due date extraction is the dominant document-level gap at 64 invoices, or 77.1% of the summary-sheet population.
Lines
Line-item completeness is split almost evenly between missing quantity (49.7%) and missing unit price (52.8%), indicating a structural extraction weakness in tabular fields.
Recon
Reconciliation fails for 17 invoices, or 20.5% of the summary-sheet invoice count.
Scope
Document KPIs come from Summary_Metrics, while distribution and vendor views use Flattened_Lines, which contains 16 distinct invoice IDs across 161 line records.
Use invoice-level charts for portfolio rates and line-level charts for pattern diagnosis.
Top vendors by issue volume
Vendor Lines Missing Qty Missing Price Failed Lines
Space Forte2424240
Go Airlines (India) Limited2424240
SNV AVIATION PRIVATE LIMITED2121210
High Furnace Manufacturer230023
ADITYA BIRLA MANAGEMENT C…6660
Aditya Birla Sun Life AMC…4440
KRISH CARS PVT. LTD.7007
Dell International Servic…12050