Continuous controls monitoring is the practice of testing internal controls over financial processes as transactions occur, across the full population, in place of a sample after the period closes. Continuous controls monitoring software reads the systems the controls run in, compares each event with the rule it should satisfy, and reports exceptions to the owner of the control.
Rever runs this as its continuous monitoring capability, with five of its fifteen agents on the job. Sentinel leads and watches unusual changes and possible control abuse. Steward watches the health of the connected systems, Auditor tests whether the approved procedures ran as written, Matcher keeps the reconciliations current, and Spend Guard reports duplicate and split spending. Each risk reaches your Inbox with its evidence and a recommended safeguard, and your team decides.
Continuous transaction monitoring and controls monitoring
The two terms describe different tests on the same data. Continuous transaction monitoring asks whether a transaction looks wrong: a bill entered twice, an amount far from the vendor's pattern, a payment to a bank account changed last week, a run of round-number claims. Controls monitoring asks whether the control around the transaction operated: did an approver with the authority sign it, did a second person review an entry the first person prepared, did the three-way match run before the payment.
A payment can pass all the transaction tests and still sit on a broken control. An approver with a lower limit can sign a correct invoice, and nothing about the invoice looks unusual. The reverse holds too: three payments that each look normal can add up to a split that keeps a purchase under an approval threshold.
| Transaction monitoring | Controls monitoring | |
|---|---|---|
| The question | Does this transaction look wrong? | Did the control around it operate? |
| A typical test | Duplicate bills, unusual amounts, payments to changed bank accounts | Approval authority, preparer and approver separation, split payments, ungated postings |
| What it needs | The transaction population | The transactions plus approvals, roles, master-data history and system logs |
| What it produces | A suspect transaction to review | A control exception with an owner and a fix |
Both depend on complete data. A test that runs on a stale bank feed or a failed sync reports a clean result it did not earn, so the monitoring has to watch its own inputs as well.
The tests a finance team starts with
A finance team adopting continuous controls monitoring tends to begin with the controls an auditor already tests on a sample:
- Approval authority: each payment and entry approved by someone whose limit covers the amount.
- Segregation of duties: the preparer of an entry or a vendor change is a different person from the approver.
- Split transactions: related payments grouped by vendor, requester and date, checked against the threshold they avoid.
- Vendor master and bank-detail changes: each change compared with its history and verified away from the request.
- Duplicates: the same bill or payment under different identifiers, including duplicate vendor records.
- Data completeness: each source feed current, and each gap reported as a gap.
Continuous controls monitoring in Rever
Sentinel leads. It groups related payments, vendors, requesters, dates and descriptions to detect splitting below approval thresholds. It tests approval authority, required sequencing, conflicting preparer and approver roles, overrides and after-the-fact approvals. It compares master details with the master change history, and for a bank-detail change it prepares an independent human callback using known-good details. Each finding states whether the concern is suspected, supported or disproved, because indicators alone do not establish fraud.
Steward watches the inputs: failed syncs, stale sources, stalled work and the log coverage of each connected system. It explains a failure in finance terms, such as a stale cash forecast or a delayed close. Steward reports a source without accessible user logs as a monitoring gap, and does not count it as a clean review.
Auditor checks that the other agents ran their approved procedures. It keeps a coverage matrix linking each procedure to a test, reperforms selected calculations and matches, and keeps explicit tests for duplicate detection, ungated hand-backs, self-approval and unauthorised hold release. Auditor reports what it tested and what it did not, and it cannot mark a material untested procedure as passed.
Matcher supplies the field-level comparisons behind unsupported-payment and duplicate findings, and Spend Guard separates a control failure from an accounting error in the duplicate and split patterns it reports.
No agent places a hold on its own. Sentinel requests a proportionate review or a payment hold, and the hold is recorded in Rever after your team approves it in the Inbox; your team applies it in the payment run. Rever records the decision with what happened next. The floors sit under all of this, in every plan: nothing leaves Rever without a recorded approval from an authorised approver, and the identity that prepares an action cannot approve it. Auditor, Sentinel and Steward come with the Growth Pro and Enterprise plans, according to Rever's pricing.
Monitoring runs beside the rest of the month's work. Continuous accounting covers the matching and the close that run beside it, and audit readiness covers the evidence your auditor reads at year end.
