Continuous Reconciliation
Reconcile continuously, not in the last week of the month
A break that appears on the third of the month is a five-minute question. The same break found on the twenty-eighth is an investigation across four weeks of activity. VapusFin matches as transactions land and puts the likely cause next to the difference.
The problem
Batch reconciliation finds the problem long after it was cheap to fix
Reconciliation is usually a period-end exercise, which means every difference is discovered at the moment there is least time to look into it. By then the context has gone: the person who posted the entry has moved on to the next thing, the bank reference has been reused, and a single mis-keyed amount looks identical to a timing difference. So breaks get written off to a suspense account, the suspense account grows, and the audit asks about it a quarter later.
What it does
Inside Reconciliation
Bank to ledger, as transactions arrive
Bank feeds are matched against ledger entries on a continuing basis rather than in a month-end batch. One-to-many and many-to-one matches are handled, along with the timing differences that make a statement and a ledger legitimately disagree for a few days.
Intercompany reconciliation across entities
Balances between entities are matched pair by pair, in both directions, across currencies. Where two entities disagree about the same transaction, the difference is quantified and attributed to one side rather than netted away.
Tax liability accounts against filed returns
Balances in tax liability and input credit accounts are reconciled against what was actually filed and what the portal shows, so a difference is found while it can still be corrected rather than during an assessment.
Sub-ledger to control account
AP and AR sub-ledgers are reconciled to their control accounts continuously, so a posting that lands in the wrong account is caught as an isolated entry rather than as an unexplained control-account difference at period end.
Breaks that arrive with a likely cause
A break is presented with the candidate explanation, not just the amount: a transposed figure, a duplicate posting, a payment recorded gross against a net receipt, a foreign exchange difference, or a transaction genuinely still in transit. The evidence for that suggestion is shown, and a person still decides.
Audit trail written as it happens
Every match, every exception, every write-off and every person who approved one is recorded at the moment it occurs, with the source records attached. The reconciliation evidence pack is a by-product of doing the work, not a project after it.
How it works
From arrival to posted entry
- 01
Connect
Bank feeds, ledger accounts, sub-ledgers and entity structures are connected, and the matching rules and tolerances for each account are set.
- 02
Match
Transactions are matched as they arrive, one to one and many to many, with tolerance rules for rounding, bank charges and foreign exchange.
- 03
Explain
Anything unmatched becomes a break with a proposed cause and the records that support it, rather than a line on a list of differences.
- 04
Resolve
Breaks route to the owner of that account. Corrections post as journals with their justification attached; write-offs need an approval that is recorded against a named person.
- 05
Close
Because matching never stopped, period end is a review of the remaining exceptions and the sign-off, not the start of the reconciliation work.
Questions
Frequently asked
- How is this different from the matching our ERP already does?
- Most ERP matching is exact and runs on a schedule. This runs continuously, matches across many-to-many relationships and tolerances, works across entities and currencies, and — where it cannot match — proposes what the difference is likely to be rather than only reporting that one exists.
- Does it post correcting entries by itself?
- Only where you configure it to, and only for the categories you allow — typically small, well-understood differences such as bank charges or rounding, within a tolerance you set. Everything else is proposed and posted after a person approves it. Every automatic posting is logged with the rule that authorised it.
- What does it need from our bank?
- A statement feed in any of the standard formats, or a direct connection where your bank supports one. Reconciliation improves with richer remittance data in the feed, but it does not depend on it — matching also works from amount, date, counterparty and reference patterns.
- Can it reconcile between entities on different ERPs?
- Yes. Entities are reconciled against each other from their own records, so a group running different systems in different countries is the normal case rather than the exception. Currency differences on intercompany balances are calculated and attributed rather than netted out.
- What match rate should we expect?
- The honest answer is that it depends heavily on the quality of the reference data in your bank feed and your ledger. We measure it on your own history during onboarding and tell you the number before you commit.
Reconcile last quarter, in a sandbox
Give us a quarter of statements and ledger extracts. We will show you what matches, what breaks, and what we think each break actually is.
Last reviewed

