{
  "title": "Purchase Invoice Processing Benchmark Data",
  "source": "Opusmatic Research Compilation",
  "context": "Dutch Technical SME / Process Industry & Installation Sector",
  "lastUpdated": "2026-07-23",
  "license": "CC BY 4.0",
  "creator": "Opusmatic",
  "invoiceExceptionHandling": {
    "metric": "Typically 20-30% of all invoices have to be treated as exceptions, resulting in very high processing costs",
    "source": "Billentis, The E-Invoicing Journey 2019-2025",
    "year": 2019,
    "url": "https://www.billentis.com/assets/reports/The_einvoicing_journey_2019-2025.pdf"
  },
  "invoiceFraudNetherlands2024": {
    "metric": "Invoice fraud reported to Fraudehelpdesk Zakelijk: 24 reports in total; among reports with quantified damage: €924,507 total, €44,024 average per incident, €25,163 median, €180,895 maximum",
    "source": "Dialogic (research for the Dutch House of Representatives), based on Fraudehelpdesk Zakelijk registrations, Table 8",
    "year": 2024,
    "url": "https://www.tweedekamer.nl/downloads/document?id=2026D29549",
    "note": "The damage statistics exclude reports with €0 or unknown damage, so the average is computed over fewer than 24 reports. Reported incidents are a lower bound; actual prevalence of invoice fraud is higher due to underreporting."
  },
  "eInvoicingAdoption2025": [
    {
      "metric": "86% of all invoices received by the Dutch central government were electronic (nearly 2 million e-invoices)",
      "source": "Digitale Overheid (Dutch government)",
      "year": 2025,
      "url": "https://www.digitaleoverheid.nl/nieuws/elektronisch-zakendoen-met-de-rijksoverheid-groeit/"
    },
    {
      "metric": "Approximately 50% of e-invoices to the Dutch central government were received via the Peppol network",
      "source": "Digitale Overheid (Dutch government)",
      "year": 2025,
      "url": "https://www.digitaleoverheid.nl/nieuws/elektronisch-zakendoen-met-de-rijksoverheid-groeit/"
    }
  ],
  "verificationArchitecture": {
    "description": "Reference architecture for controlled automated invoice processing",
    "controls": [
      {
        "control": "Per-supplier checksum reconciliation",
        "purpose": "Verifies extraction completeness and correctness by comparing the sum of extracted lines against independently extracted per-supplier subtotals; a mismatch pinpoints the specific supplier block to review."
      },
      {
        "control": "Grand-total reconciliation",
        "purpose": "Hard pass/fail gate: the sum of all extracted lines must equal the report's independently extracted grand total."
      },
      {
        "control": "Invoice matching",
        "purpose": "Each payment line is approved only when supplier, bank account number, and amount match an independently processed, actually received invoice; catches duplicates, ghost invoices, and manipulated account numbers."
      },
      {
        "control": "Human-in-the-loop exception handling",
        "purpose": "Any line failing verification is surfaced to an employee for review; nothing is paid silently."
      }
    ]
  }
}