The problem
Intercompany reconciliation is one of the most painful and time-consuming parts of month-end close. Finance teams typically pull trial balances, AR and AP listings from several ERP instances or entity ledgers, drop them into spreadsheets, and then manually match transactions across entities. Currency differences, timing differences, mis-postings and inconsistent reference data make every cycle feel slightly different. By the time mismatches are identified, the close clock is already against the team.
The process is usually held together by a small number of experienced people, a shared drive of evolving templates, and email chains chasing counterparties inside the same group.
Why it matters
Unresolved intercompany differences distort consolidated results, delay close, create audit findings and can lead to tax exposure. They also consume senior finance time that should be spent on commentary, analysis and decision support. As groups grow through acquisition, the number of entity pairs grows quickly, and the spreadsheet-based approach simply does not scale.
From a control perspective, manual matching with limited evidence trail is a recurring concern for auditors, particularly where adjustments are posted late in the cycle without clear approval.
The opportunity
A governed, no-code workflow can connect directly to source ledgers, standardise the data, apply consistent matching logic, and surface only the genuine exceptions that need human judgement. AI can support the judgement layer — suggesting likely matches for fuzzy descriptions, classifying the nature of differences, and drafting clear commentary for review notes.
The goal is not to remove the accountant from the process. It is to remove the mechanical work so the accountant can focus on resolving the differences that actually matter.
Example workflow
1. Connect the source data
Pull AR, AP, intercompany loan and recharge data directly from each ERP or entity ledger. Include trial balance extracts for the intercompany account ranges. Where direct connections are not possible, use governed file drops with schema validation.
2. Standardise and prepare the data
Normalise entity codes, currency, transaction dates, document references and counterparty identifiers. Translate balances to a common reporting currency using the agreed rate table. Flag missing or malformed records before they enter the matching engine.
3. Apply business logic
Run a tiered matching routine:
- Exact match on reference and amount
- Match on reference with FX or rounding tolerance
- Fuzzy match on description, date window and amount
- AI-assisted suggestion for residual items
Each match is tagged with the rule that produced it, so the logic is fully transparent.
4. Run checks and controls
Reconcile matched totals back to the trial balance for each entity pair. Check that every transaction is either matched, classified as a timing difference, or routed to an exception queue. Apply materiality thresholds and confirm that no entity pair is missing from the run.
5. Produce outputs
Generate an intercompany matrix showing balances by entity pair, matched and unmatched amounts, timing differences and net exposure. Produce supporting schedules for each pair and a summary pack for the group controller.
6. Review exceptions
Unmatched items are routed to the responsible accountants in each entity with context, suggested counterparty, and a place to add commentary or upload supporting evidence. AI can draft a first-pass explanation that the accountant reviews and edits.
7. Move to governed operation
Lock the workflow with version control, role-based access and a full audit log of inputs, rules applied, matches made and adjustments posted. Schedule the run on a defined close calendar with sign-off at controller level.
What good looks like
- Source data pulled directly from ledgers with no manual re-keying
- Consistent matching rules applied across all entity pairs
- Clear separation between matched, timing and genuine differences
- Exceptions routed to named owners with deadlines
- Full audit trail from raw data to posted adjustment
- AI used to assist judgement, never to post entries unsupervised
- Repeatable run that produces the same answer given the same inputs
Benefits
For the finance team
Less spreadsheet wrangling, fewer late nights at close, and a clearer view of which differences actually need attention. Knowledge is captured in the workflow rather than sitting with one or two people.
For leadership
Faster, more reliable close. Cleaner consolidated numbers. Reduced audit friction and lower risk of restatement. A defensible, documented process that scales as the group grows.
For the wider business
Entity finance teams get earlier visibility of issues affecting their books, and a consistent way to respond. Tax and treasury get cleaner data to work from.
Where to start
Pick the two or three entity pairs that cause the most pain at close. Map the current process honestly, including the spreadsheets, the email chains and the workarounds. Build the first version of the workflow against those pairs, prove the matching logic, and then extend coverage entity by entity. A focused pilot usually delivers visible benefit within one or two close cycles.
How 4th Revolution can help
4th Revolution is finance-led. We combine practical accounting knowledge with data engineering, no-code automation and embedded AI. We do not just build a workflow and walk away — we help you create a governed, repeatable process that your team owns, your auditors trust, and your controller can rely on at close.
We focus on the controls, the evidence trail and the handover, not just the technology.
Example outcome
Before: a group finance team spends the first five working days of close on intercompany matching across twelve entity pairs, working from exports into a shared spreadsheet, with late adjustments posted on day six.
After: source data is pulled automatically on day one, matching runs overnight, and the team starts day two with a clear exception list, owners assigned and AI-drafted commentary ready for review. Close completes earlier, with a documented audit trail and materially fewer unresolved differences carried forward.