Intercompany Matching and Settlement Should Start Before Reconciliation

Intercompany Matching and Settlement Should Start Before Reconciliation

At period-end, the controller often inherits an avoidable choice: delay the close while teams investigate mismatches, or accept adjustments supported by incomplete evidence. Intercompany matching and settlement is usually treated as a reconciliation problem. In practice, many exceptions can originate earlier—when counterparties record the same transaction under different identifiers, periods, currencies, or policy interpretations. The controller’s priority should therefore be to validate transaction authority and context before balances reach reconciliation.

This matters as transaction reporting becomes more time-sensitive. The EU’s adopted VAT in the Digital Age rules introduce real-time, e-invoice-based reporting for qualifying cross-border B2B transactions from July 1, 2030. Not every intercompany balance falls within that scope, but applicable supplies will require structured, reliable transaction data earlier in the process (European Commission ViDA overview). Organizations already struggle with the inputs: in a self-selected Deloitte webcast poll of 4,114 finance, accounting, and tax participants, nearly 60% cited data quality and reconciliation or transfer-pricing documentation as their greatest intercompany tax-management challenges (Deloitte intercompany accounting analysis).

The exception queue is hiding a source-authority problem

The general ledger and subledgers provide structured entries, but safe matching may also depend on semi-structured invoices and settlement instructions, plus unstructured email explanations and approval history. Entity identifiers may differ between systems; transaction references may be missing; one party may book gross while another books net; or a late adjustment may lack its original policy rationale.

This is a mixed-data problem because structured accounting records must be reconciled with semi-structured documents and unstructured exception context before a posting or settlement decision is safe to execute. When that context remains outside the workflow, matching logic produces false exceptions, teams duplicate investigations, and approvals become difficult to audit.

Intercompany matching fails less from an inability to compare balances than from an inability to establish which evidence governs the transaction.

Cycle time provides a useful diagnostic. APQC reports a cross-industry median of 2.0 calendar days to process an intercompany transaction for posting and reconciliation, based on 522 companies. The public benchmark does not cover every upstream approval or downstream cash-settlement step, and it is not a best-in-class target (APQC intercompany cycle-time benchmark). Controllers should use it to question where their own elapsed time sits: in transaction processing, evidence collection, exception ownership, or approval.

Move validation upstream, then automate selectively

The target operating model should create one governed path from transaction evidence to system-ready output.

1. Capture and normalize. Bring together counterparty entries, entity master data, invoices, currencies, fiscal periods, settlement instructions, and available approval evidence. Map them to a common transaction schema.
2. Validate and reconcile. Check required fields, entity relationships, currency treatment, duplicate references, and mirror-entry consistency before applying approved matching rules.
3. Route genuine exceptions. Send unresolved matches to the responsible accounting, tax, or treasury owner, with the conflicting fields and supporting evidence attached.
4. Approve and deliver. Produce approved postings, adjustments, or settlement instructions in the required downstream-system schema, retaining the decision trail.

StarDox Intelligence can support that model as an enterprise transaction-automation and decision-support layer for validation and exception handling, without replacing the ERP, accounting policy, or required judgment.

Automation should stop when a proposed match has a missing mirror entry, conflicting entity or currency authority, or exceeds the organization’s financial-materiality threshold. The controller or delegated policy owner should receive both entries, source documents, validation results, and prior approvals; the permitted next action is to approve a corrected posting or settlement, request additional evidence, or reject the match.

Three actions for the next close

Improvement should begin with a controlled baseline rather than a broad technology program:

1. Map source authority for entity, currency, period, pricing, and settlement fields, including who resolves conflicts between records.
2. Baseline exception performance by measuring initiation-to-reconciliation cycle time, first-time-right matching, exception volume, and aging by root cause.
3. Define approval boundaries for low-confidence, policy-sensitive, and material transactions, then test those boundaries through one close as a workflow diagnostic.

Sources

• Deloitte: Managing intercompany accounting challenges and tax complexities
• APQC: Average cycle time to process an intercompany transaction
• European Commission: VAT in the Digital Age