Bank statement analysis built for underwriting, not just extraction.
The true cash flow of a business, from its statements: money in, money out, debt service against revenue, the cash buffer, and whether the statement itself is consistent. Anything unclear is flagged rather than guessed.
- PDF, scanned or CSV
- Every account netted
- Flagged, not guessed
- Statement consistency checked
- Hosted service, in your environment, or as an API
- Osko receipt, Footscray Tiling+A$4,200Revenue
- BPAY to the tax office-A$3,100Tax
- Square settlement+A$1,860Revenue
- Transfer from the owner+A$5,000Owner money in
- Weekly debit, finance provider-A$1,150Lender repayment
- Transfer to business savings account-A$2,000Own-account transfer
Sound familiar?
- "Transfer in"A sale, a loan or the owner's own money? The statement does not say, and the answer changes the file.
- Revenue that looks healthyUntil a loan is found sitting in it, counted as sales.
- Hidden lendersRepayments to other lenders buried in "transfers out", so the stacking only shows up after settlement.
- Money counted twiceSeveral accounts, and every move between them looks like new income.
- One rule for every businessA builder's uneven receipts and a cafe's thin buffer are judged by the same rule, so the wrong ones get flagged.
Not a statement extractor
Decisions you can defend.
What the money shows
A one page summary for the underwriter or broker, with every line sourced: revenue classification, lenders and stacking, transfers and the owner's money, dishonours and risk tracking, trends against the business's own normal, rhythm, obligations and debt service.
Built for messy statements and for each business's own rhythm
Digital and scanned PDFs, CSV and XLSX exports from any bank, across several accounts, with balances checked and unclear lines flagged rather than guessed. Each business is read against its own history and the same months a year earlier, so a quiet season is not read as a decline. Revenue, lender repayments, owner drawings, tax and super, supplier terms and customer concentration are read as a business's money, not as spending categories.
Where it goes
- ServiceabilityWorked out under your policy.
- The underwriting memoCash flow, income, expenses and commitments flow into the memo, each with its source.
- Credit models and early warningBehavioural signals for scoring and for watching the book, including the cashflow score.
- BrokersA profile ready to submit, built from the statements alone when that is all there is.
- An APIFor lenders, brokers, accountants and the platforms they use.
Underwriting analysis, extraction-only tools and manual spreading compared
| Underwriting analysis | Extraction-only tools | Manual spreading | |
|---|---|---|---|
| Labels by economic role | Every line, with a reason | Raw rows, or merchant categories | By eye |
| Transfers and own accounts | Netted across the submitted accounts | Left in as income and spend | Checked by hand |
| Existing lenders and stacking | Found, with repayment cadence | Usually not identified | Easy to miss |
| Balance reconciliation | Checked, with gaps flagged | Not always checked | Occasionally |
| Comparison with the business's own normal | Latest month and same months a year earlier | Not provided | Rarely done |
| Output | A one page summary and memo inputs | A spreadsheet | Notes |
| Volume | Built to run on every file | Built for volume, returns rows | Limited by people |
Questions about statement analysis
A deposit lands. Sale, loan drawdown, refund or the owner's own money?
It is classified from the description, the recurring pattern, the business type, the other accounts submitted and the supporting documents, so transfers and loan proceeds never count as sales.
The business runs three accounts. How is the same money not counted twice?
Accounts are analysed together in one request. Transfers between the business's own accounts are netted, and the summary is at business level.
What shows up that a transaction export does not?
An export gives you rows. We interpret what each row means in a lending context: a deposit may be a sale, a loan drawdown, a refund or the owner's own money, and the answer changes the file. Unclear lines are flagged for review, not guessed.
How do we see the lenders a business already has?
Recurring lender debits are matched by cadence and funder pattern, so weekly and daily repayments show the lenders already in place, and new ones appearing close together are flagged as stacking.
What is on the one page?
Monthly revenue and trend, deposit consistency and customer concentration, average daily balance, overdrawn and negative-balance days, dishonours, buffer in weeks of outgoings, debt service against revenue, owner drawings, payroll, superannuation and tax on their usual dates, and the business against its own history.
The statements are scans and phone photos. Does that still work?
Digital and scanned PDFs, CSV and spreadsheet exports from any bank go into one schema. Balances are checked line by line, and anything that does not reconcile is shown.
Why is a cafe's quiet month not a red flag?
Labels and norms adapt to the business type and its operating rhythm. Each business is compared with its own history and the same months a year earlier before it is compared with anyone else, so a cafe's quiet month and a tradesperson's lumpy receipts are not false alarms.
What happens to a line you cannot be sure about?
It is flagged for review with the reason, never guessed, and counted in the needs-review share so you can see how much of a file rests on judgement.
Where do the results end up?
Into the underwriting memo, the Cashflow Score, serviceability under your policy, your LOS or CRM, or the API.
Can we use it without the memo?
Yes. It runs as an API or a file batch for lenders, brokers, accountants and platforms, and as the evaluation on your own statements.
What you can buy
Start a pilot: bank statement analysis
Who it is for
Business lenders, broker aggregators, and lending and accounting platforms that want an API. Credit analysts and the people who manage them.
What you get
- Every line of every account labelled by what it is to the business; transfers between the business's own accounts netted; balances reconciled line by line
- Anything unclear flagged for a person with the reason, never guessed
- A one-page summary per business: what it earns, who it already owes, what the owner takes out, and whether the statements add up
- An API for one file or a file batch for a book; PDF, scanned PDF, CSV and XLSX from any bank
- The pilot report: reconciliation rate, needs-review rate and agreement with your own analysts' labels
Cashflow signals, read the way the business trades
Revenue, lender repayments, owner drawings, tax and super, supplier terms and customer concentration are read as a business's money, not as spending categories. The cashflow signals are engineered for business lending and for each business type, and read the business's operating rhythm: seasonality, payment cycles, weekday and month-end patterns. Each business is compared with its own history and the same months a year earlier before it is compared with anyone else.
How it arrives
A sandbox in your environment on a few of your own files while your privacy and security review runs, then the hosted service in Sydney: files in by upload or API, results back, nothing to install. Institutions whose policy requires it run it in their own environment.
The first step
A fixed-fee pilot on statements you have already analysed, with success criteria agreed first, measured against your own analysts. The winner is named even when it is not us.
- Where it is used
- Banks and lenders, Brokers and aggregators
- One case
- Revenue overstated by inflows (Illustrative worked example)
See what a business really earns.
Start with statements you have already analysed.