Fix the rules before you add AI. Once every entity books intercompany the same way, AI can match both sides across ledgers, explain each difference in plain words and draft the query to the other entity's accountant. A group controller approves every proposed adjustment, and nothing posts to the ledger without that approval.
#Key takeaways
- Intercompany breaks are mostly a process problem: different cut-offs, currencies, account mappings and invoices booked on one side only.
- AI helps with the tedious middle: matching transactions that do not share a reference, grouping differences by cause and chasing the other entity.
- A controller keeps the rules, the eliminations and every adjusting entry.
- If your entities share one ERP instance, its built-in intercompany module may be enough once the rules are agreed.
- Start with the pair of entities that causes the most breaks, not the whole group.
#What does intercompany look like at month-end for a group controller?
A group with a holding company, a trading company in Dubai, a branch in Riyadh and a service company somewhere else books recharges, management charges, loans and shared costs between them. Each entity closes on its own ledger, sometimes on different systems: SAP at the parent, Odoo or QuickBooks at the smaller entities. Then the controller exports both sides into Excel and starts matching.
The breaks are familiar: one side booked the recharge, the other has not; amounts differ by an exchange rate; a VAT line was split differently. None of this is new. lucanet, citing BPM International, checked november 2026 reports that 72% of surveyed companies spend significant time clearing intercompany differences, and the SAPinsider, december 2023 financial close study found 55% of SAP customers called intercompany accounting a significant and growing challenge.
#What can AI do in intercompany reconciliation?
- Match transactions between two entities even when references, descriptions or languages differ, using amount, date, counterparty and invoice text.
- Group unmatched items by likely cause: timing, exchange rate, missing booking, wrong account, VAT treatment.
- Draft the query to the counterparty accountant with both sides of the transaction attached.
- Prepare a proposed adjusting entry for the controller to review, with the evidence beside it.
- Summarise the open differences per entity pair before the consolidation meeting.
What stays human: the group's intercompany policy, the agreed exchange rate source, which side corrects, the eliminations, and every entry posted. The controller signs off; the AI prepares the work.
#What does a practical intercompany playbook look like?
- Write one intercompany policy: which transactions, which accounts, who invoices, which rate, which cut-off.
- Map each entity's intercompany accounts to one shared list.
- Agree a settlement rule so balances do not build up between entities.
- Pick the entity pair with the most breaks and run the AI matching beside your manual work for one close.
- Measure the unmatched items at first pass, the items cleared before consolidation, and the adjustments the controller rejected.
#When is the ERP's own intercompany module enough?
If all entities sit in one SAP, Dynamics 365 or NetSuite instance and book intercompany through the system's own module, fix the policy and the account mapping first. That alone removes most breaks. A consolidation tool may also be the right buy for a large group. AI matching earns its place when entities run on different systems, documents arrive in Arabic and English, or the module cannot read free-text invoices.
#How does Stennir approach intercompany for a group?
Stennir usually starts with AI consulting, because intercompany pain is often a policy gap rather than a tooling gap. The AI Opportunity Audit maps your entities, systems and breaks, and returns five ranked workflows, each with a cost and a payback case, plus one recommended build with its scope and hosting region. The report is yours whether or not you build with Stennir.
If the build is intercompany matching, AI implementation runs it across your ledgers with a controller approving every entry and every action logged. AI training helps the entity accountants use it.
We are
Stennir starts with your intercompany rules and builds matching across your existing ledgers, with a controller approving every adjustment.
We aren't
Stennir does not replace your consolidation tool, and does not post or eliminate entries on its own.
#What should a group controller do next?
List your entity pairs and count the unmatched items at last close. If one pair holds most of them, start there. Read more at AI for finance and in AI for bank reconciliation, then take the free readiness assessment or book a 30-minute call.
#Common questions
Do all our entities need to be on the same ERP?
No. The matching reads exports or connections from each ledger, so a parent on SAP and subsidiaries on Odoo or QuickBooks can be reconciled together. What does need to be shared is the intercompany policy and the account mapping, which the audit helps you agree before anything is built.
What about different currencies?
The workflow applies the exchange rate source your group policy names, then shows any remaining difference as a rate difference rather than a missing booking. The controller decides how to treat it. The AI never picks a rate on its own or posts a revaluation entry without approval.
What do you need from us?
Exports of intercompany accounts from each entity for a recent close, your current policy if you have one, the account mapping, and one person per entity who answers queries. The group controller is named as the approver for every proposed adjustment.
Where does group data stay?
In your own accounts and the hosting region your rules require, which matters when entities sit in different countries. Stennir does not copy ledgers into its own systems. You keep the accounts, the data and the off switch, and each entity's access can be limited to its own records.
