Skip to content
All blog
AI Accounting Group

Intercompany transactions without the spreadsheet

David 7 min read

Every group with more than one entity has the same conversation at period end. Company A says Company B owes it RM 84,300. Company B's books say RM 79,150. Nobody knows which is right, and somebody spends two days finding out.

The difference is almost never fraud or incompetence. It is structural: two entities record the same transaction independently, at different times, from different documents.

That is a design problem, and it is fixable.

Why the balances diverge

Timing. A recharges B on the 29th. B receives the document on the 3rd and posts it in the following month. Both correct in their own books, and the balances disagree at period end.

Different amounts. A raises a recharge for RM 12,000. B queries part of it, agrees RM 10,500, and posts that. A never amends its side.

One-sided entries. A posts a management charge and tells nobody. It appears on A's books and nowhere on B's until someone reconciles.

Different classification. A treats it as a recharge of costs; B treats it as a service purchase. Both defensible, and consolidation now has two different things to eliminate.

Currency. Both entities record the same amount in different currencies at different rates, so the balances cannot agree by construction until someone decides which rate governs.

The structural fix

The important shift: stop recording intercompany transactions twice.

If the transaction is created once and posts to both entities simultaneously, timing differences disappear entirely — not reduced, eliminated. Both sides carry the same amount, the same date, the same reference, and the same classification, because there is one record rather than two.

That single change removes the first three causes above. It is a bigger improvement than any amount of faster reconciliation, because reconciliation is the cure and this is prevention.

Where entities genuinely must post independently — different systems, different jurisdictions with local requirements — the fallback is matched references and automated reconciliation on a shared identifier, which reduces the problem without eliminating it.

What automation adds

Continuous matching. Intercompany balances reconciled as transactions occur rather than at period end. A difference discovered on the day it arises is a two-minute conversation; the same difference at year end is an investigation.

Difference classification. Not "the balances differ by RM 5,150" but which specific transactions are unmatched and why — in transit, amount disputed, one-sided, classification mismatch.

Automatic elimination. Consolidation entries derived from matched pairs rather than assembled manually. This is where manual group accounting consumes the most time and produces the most errors.

Ageing. An in-transit item is normal this month and a problem next month. Ageing the differences distinguishes them.

The control that catches what is left

Even with shared records, reconcile intercompany balances monthly and confirm they agree to zero on elimination.

The specific thing to watch: a persistent difference that everyone has stopped questioning. Groups accumulate these. There is a RM 3,400 difference that has been there for two years, it has been carried forward every period, and nobody remembers what it was. Each period it is too small to investigate; cumulatively it is a hole in the consolidation and it will be asked about eventually.

Set a rule that any intercompany difference older than a defined period gets investigated regardless of size. It is the only way these get resolved, because on any individual month the rational choice is always to leave it.

Transfer pricing sits on top of this

Worth separating, because it is frequently confused with the reconciliation problem.

Whether intercompany charges are at an appropriate arm's-length basis is a policy and documentation question, not a bookkeeping one. Automation ensures the charges are recorded consistently and that the basis is attached to the transaction — which makes the documentation possible — but it does not determine what the charge should be.

The practical contribution: if every intercompany charge carries its basis and calculation, assembling the supporting documentation stops being an archaeology exercise.

What still needs a person

  • Deciding the basis on which entities charge each other
  • Resolving a genuine dispute between two entities about what was agreed
  • Judgement on classification where both treatments are defensible
  • Anything involving different jurisdictions' requirements

Common questions

Why do intercompany balances never agree?

Because two entities record the same transaction independently, at different times and from different documents, so timing differences, amended amounts, one-sided entries and inconsistent classification all accumulate. The structural fix is creating the transaction once so it posts to both entities simultaneously, which eliminates those causes rather than making the reconciliation faster.

Can intercompany reconciliation be automated?

The matching and elimination can be, and continuous matching is more valuable than period-end matching because a difference found on the day it arises is a short conversation rather than an investigation. Automation is most useful when it classifies differences by cause — in transit, amount disputed, one-sided, classification mismatch — since each requires a different response.

What is the most common hidden problem in group accounts?

A small persistent intercompany difference that has been carried forward for years and that nobody remembers the origin of. On any individual month it is too small to justify investigating, so it never gets resolved, and cumulatively it is an unexplained hole in the consolidation. A rule that any difference older than a set period gets investigated regardless of size is the practical remedy.

Does automation handle transfer pricing?

No. Whether intercompany charges are at an appropriate arm's-length basis is a policy and documentation matter rather than a bookkeeping one. What automation contributes is ensuring charges are recorded consistently with their basis and calculation attached, which makes assembling the supporting documentation straightforward instead of an archaeology exercise.


Related: AP automation for a multi-entity group · ai accounting for a group of companies · closing the books with AI assistance


See what you could build

Start a free trial and describe what your business needs in plain language — SmartB Studio builds the module for you.

Start free trial
Get started

No credit card · Cancel anytime · Your data stays yours