What makes agency accounting its own discipline
An insurance agency handles two kinds of money that must never touch. Premium belongs to the carrier and to the insured — it passes through the agency and is held in a fiduciary capacity. Commission is the agency's own revenue, earned on business placed, and it arrives on somebody else's schedule.
Most jurisdictions require the first to be held separately from operating funds, and commingling the two is one of the fastest routes to a licensing problem. The rules and penalties vary state by state; California, for example, requires a monthly three-way reconciliation of the bank statement, the trust ledger and the client or carrier sub-ledger.
Underneath the regulation is a mundane data problem: the bank statements arrive as PDFs, the carrier statements arrive as different PDFs in a dozen different layouts, and the agency management system holds a third version of the truth. Getting all three into comparable data is most of the work.
Premium is fiduciary money, not cash flow
When a client pays a premium to the agency, that money is held for the carrier. It is not working capital, it is not revenue, and drawing on it to cover payroll in a slow month is not a cash-flow decision — in most states it is a violation with real consequences.
The practical requirement is a separate premium trust account, with commission drawn out only when it has genuinely been earned and with documentation supporting each withdrawal. Regulators look for an audit trail showing which amount was commission and when it became the agency's, not a monthly sweep of whatever looked spare.
Our part is narrow and useful: the trust account statement as complete, provable data so the balance can be compared to what the agency owes carriers and clients at that date. We do not perform the reconciliation, hold a sub-ledger, or advise on your state's code — that belongs with your accountant and your state's insurance department.
Agency bill and direct bill are different accounting problems
Under agency bill, the agency invoices the insured, collects the premium into trust, deducts commission and remits the balance to the carrier. Cash and commission move together, the trust account does real work, and the reconciliation is about what is still owed to whom.
Under direct bill, the carrier collects from the insured directly and pays commission to the agency later, based on policy activity. Commission does not move with premium cash at all — which is precisely why a direct-bill commission statement never lines up with a bank deposit without effort.
Most agencies run both, often with the same carrier, and the mistake is treating them with one process. The trust account matters enormously for one and barely at all for the other; the reconciliation burden inverts. Deciding which lines belong to which model before you start is what makes the month close.
| Agency bill | Direct bill | |
|---|---|---|
| Who collects premium | The agency, into trust | The carrier |
| Where commission comes from | Deducted from premium held | Paid later by the carrier |
| Trust account involvement | Central | Minimal |
| Reconciliation question | What do we still owe the carrier? | Were we paid what we earned? |
| Typical failure | Commission drawn before earned | Statement never reconciled to the deposit |
| Document to extract | Trust statement + carrier account current | Commission statement + bank deposit |
The commission statement never matches the deposit
A carrier sends a commission statement listing dozens or hundreds of policy-level entries: new business, renewals, endorsements, cancellations, adjustments. Then it sends money — one deposit, for the net, on its own schedule, often days later and occasionally combined with something else.
So the reconciliation is not line to line, it is statement total to deposit, and then policy by policy inside it when the totals disagree. Doing that from two PDFs is exactly as unpleasant as it sounds, which is why a great many agencies simply do not do it and accept whatever the carrier says they earned.
That acceptance is expensive. Missing renewals, commission paid at the wrong rate and policies that never appeared on any statement are all common, and none of them announce themselves. Having both sides as data turns an unpleasant afternoon into a sort exercise where the differences fall out.
Every carrier has its own statement format
An agency of any size deals with fifteen to forty carriers, and every one of them formats its commission statement differently. Different column names, different treatment of negatives, different ways of showing a chargeback, some in landscape, some as a scan of a printout, a few with the totals on a separate page.
That variety is exactly what defeats template-based tooling and hand-built parsers: the moment you have a rule per carrier, you have a maintenance job per carrier. It is the same problem invoice processing has with suppliers, at a smaller scale but with higher stakes per document.
Extraction reads them without per-carrier setup, and merging normalises the result — the same columns across every carrier, with a column identifying which one each row came from. That normalised sheet is what makes any cross-carrier question answerable at all.
Chargebacks arrive quietly and land on producers
When a policy cancels or lapses early, the carrier claws back commission it already paid. Each carrier has its own chargeback schedule and the rules differ by product line, so the same cancellation can produce very different outcomes across a book of business.
The accounting problem is that a chargeback usually appears as a negative line inside a much larger statement and is netted out of the deposit. Nobody sees it as an event; it just makes a month's income slightly lower than expected, which is easy to explain away and hard to trace later.
Where it really bites is producer compensation. If a producer was paid a split on commission that has since been clawed back, the agency has an internal recovery question that has to be based on evidence. Extracted commission data with negatives preserved as negatives is what makes that conversation short.
The differences worth investigating, and the ones that are not
When a carrier statement total does not equal the deposit, the cause is nearly always one of a short list. Recognising which one you are looking at decides whether it needs an email to the carrier or simply a note in the file.
Timing differences resolve themselves; the others do not. The ones in the lower half of the table are where agencies routinely lose money quietly, because nothing about them looks unusual on a single statement.
| Difference | Cause | Action |
|---|---|---|
| Deposit later than the statement | Carrier payment cycle | Match on statement period, not deposit date |
| Two statements, one deposit | Carrier combined a payment run | Match combined totals for the period |
| Deposit lower than statement | Chargebacks netted off | Identify the negatives and allocate them |
| Policy missing from the statement | Not credited to the agency, or a carrier system change | Query the carrier with the policy number |
| Rate lower than expected | Commission schedule changed or applied wrongly | Check against the carrier agreement |
| Renewal never appeared | Book moved, lapsed, or credited elsewhere | Query — this is the expensive one |
| Unexpected lump sum | Contingent or profit-sharing payment | Tag separately from ordinary commission |
Producer splits need a reliable revenue number
Producers are usually paid a share of commission, sometimes at different rates for new business and renewals, sometimes with house accounts excluded. All of that arithmetic depends on one input: what the agency actually received, per policy, in the period.
If that input comes from expectation rather than from carrier statements and bank data, the splits drift. It is rarely dramatic — a renewal that never paid, a rate that changed, a chargeback nobody allocated — but it compounds, and producers notice long before the agency does.
We do not calculate splits, hold producer agreements or run payroll. What we provide is the received-commission dataset those calculations should rest on, at policy-line granularity where the carrier statement gives it.
Contingent and profit-sharing payments distort a month
Contingent commissions, profit-sharing and volume bonuses arrive as occasional lump sums, often in the first quarter, based on the prior year's loss ratio and volume. In the bank they look like a very good month.
Treating them as ordinary revenue makes every trend line useless — one quarter looks spectacular and the next looks like a collapse. Tagging them separately in the data is a two-minute discipline that makes a year's figures honest.
They also usually relate to a year in which producer arrangements may have differed, so allocating them requires knowing what they are for. Keep the carrier's letter or statement with the deposit; the amount alone tells you nothing a year later.
The three-way check, where it applies
Where a state requires it — California is the clearest example — the monthly check compares the trust bank statement, the agency's trust ledger and the sum of client or carrier sub-ledger balances, at one cut-off, to the cent. Any difference means money is not where the records say it is.
The same three-legged shape appears in law firms and property management, and the failure mode is identical: the bank leg is often typed rather than imported, so a difference might be a real problem or might be a transposed digit, and an hour disappears finding out which.
Extraction removes that ambiguity from the bank leg. Every line as data, with a completeness proof against the statement's own opening and closing balances, so a difference is a genuine reconciling item by the time anybody investigates it.
Why a dropped line matters more in trust accounting
In ordinary business accounting, a missing transaction is an inconvenience. In fiduciary accounting it undermines the only control that proves you are holding what you should be holding — and it does so silently, because a listing with one row missing looks completely normal.
Every statement we extract is checked against its own arithmetic: opening balance plus the transactions must equal the closing balance. When it reconciles, the listing is provably complete; when it does not, the document is flagged rather than passed along quietly.
The cross-month check matters too: each month's closing balance should equal the next month's opening. That catches the missing statement, which in agencies is usually the month somebody changed banks or the account was renamed after an acquisition.
Where this sits beside an agency management system
An agency management system holds the policies, the clients, the producers, the commission expectations and — where carriers support it — the direct download of commission data. That is the right home for all of it and this replaces none of it.
The gap is what the AMS cannot pull. Carriers that do not support download, smaller markets and MGAs that email a PDF, historical periods before the AMS was implemented, statements from a book of business you just acquired, and the bank side of accounts with no feed.
Where a carrier offers a genuine electronic feed into your AMS, use it — it comes from the source and needs no reading. Conversion is for the long tail, which in this industry is stubbornly large and shows no sign of disappearing.
Buying a book of business is a data problem
Agency acquisitions are common, and diligence rests on verifying revenue that the seller reports. The seller's system says one thing; the carrier statements and bank deposits say what actually arrived. Reconciling the two is the single most useful piece of financial diligence available.
That work is document-heavy: a year or two of commission statements from every carrier, plus the bank statements that received them, converted and normalised so revenue can be totalled by carrier, by line and by month. Retention and chargeback patterns show up in the same dataset.
After completion, the same normalised history is what makes the first year of ownership legible — you know what the book earned before you touched it, which is the baseline every later argument depends on.
What an examination asks for
A state examination or a trust account audit asks for the statements, the reconciliations, the sub-ledger listing at the same dates, and support for withdrawals of commission from the trust account. The theme is documentation rather than sophistication.
The most common failing is not fraud, it is untidiness: a commission draw with no supporting calculation, a reconciliation done but not retained, a missing month of statements. A folder per month with the original statements, the converted workbook and the signed worksheet answers most of it before anyone asks.
Since originals are deleted immediately after extraction on our side, that folder has to live in your storage — the extracted data is the working record, the bank's and the carrier's documents are the evidence.
How this relates to our other pages
This is a vertical page about an agency's own money. If your interest is the underlying documents rather than the industry, bank statement to Excel and statement of account to Excel cover the general cases.
Where an agency also handles significant supplier invoices or wants the payables side, accounts payable automation and duplicate payment detection apply the same way they do to any business.
And if you are building this into software rather than doing it in a spreadsheet, the document extraction API and the guide to choosing a parsing API are the developer-facing versions.
The monthly workflow
The sequencing matters: proving the bank data first means that when a total does not match, the investigation starts at the carrier rather than at your own typing.
1. Collect
Trust statement, operating statement, and every carrier commission statement for the period.
2. Convert
All of them at once. Carrier statements normalise into the same columns with a carrier column added.
3. Prove
Each bank statement checked against its own opening and closing balances before anything downstream.
4. Match totals
Each carrier statement total against the deposit that carried it, allowing for timing.
5. Investigate differences
Only where totals disagree, drop to policy level — that is where missing renewals and rate errors live.
6. Reconcile trust
Bank, ledger and sub-ledger at the same cut-off, where your state requires it.
7. File it
Statements, workbook and worksheet in one folder per month, in your own storage.
The boundary, stated plainly
FlowParse is a document-extraction engine, not an agency management system and not accounting software. There is no policy record, no client file, no commission ledger, no producer agreement, no carrier download and no trust sub-ledger.
It does not verify commission against your carrier contracts, calculate producer splits, decide when commission is earned, or perform the three-way reconciliation — two of those legs live in systems we never see. It also gives no advice on your state's insurance code, trust account rules or licensing obligations; those belong with your accountant, your counsel and your state department of insurance, and they differ substantially by state.
And it is not an archive: originals are deleted immediately after extraction. The statements an examination will ask for must live in your own storage.
A worked example: 22 carriers, one afternoon a month
A twelve-person property and casualty agency worked with twenty-two carriers, about two-thirds direct bill. Commission statements were filed in a shared drive and, in practice, reconciled once a year if at all — the format variety made monthly checking feel impossible.
The change was to convert every commission statement and both bank statements in one pass each month, normalising the carrier data into a single sheet, then compare statement totals to deposits. Investigation only happened where a total disagreed.
In the first three months the exercise found a renewal book from one carrier that had stopped appearing on statements after a system migration on the carrier's side, and a run of chargebacks that had never been allocated back to the producers involved. Neither was visible in the agency management system, because the system held what was expected rather than what arrived.
Where to go next
For the conversion mechanics see bank statement to Excel; for many documents at once see batch processing; for normalising many carriers into one sheet see consolidate bank statements and merge PDF to Excel.
For proving completeness, bank statement validation. For the commission statement itself as a document type, statement of account to Excel and remittance advice to Excel cover the closest general cases.
Adjacent verticals with the same client-money shape: property management and law firms.
Find out what you were actually paid
Convert every carrier statement and both bank accounts for the month, then compare totals — the differences are where the missing commission is.