Financial evidence review
Dealensics cross-checks bank statements, BAS and financial reports against the figures they are meant to support — and shows what reconciles, what doesn’t and what’s missing.
For buyers of a business and their advisers
From a sample case
AUD30,000
not reconciled · 21.4% of the seller’s claim
For people separating and their advisers
From a sample case
92 days
with no statements supplied for one savings account
For brokers and business borrowers
From the NORTHSTAR-01 design case
7 requests
each naming the document, the account and the dates
Who it’s for
Dealensics is for the people who have to rely on a stack of financial records — and for the advisers who help them. Three modules, one method.
Buying a small or medium-sized business and want to see what the seller’s numbers rest on.
Acquire · Early accessPreparing a client’s due diligence and want the evidence organised and the questions drafted.
Separation · In developmentWant a calm, factual picture of the household’s accounts, values and records.
Separation · In developmentNeed the financial records organised, dated and comparable before they advise.
Credit · In developmentPreparing a lending file and want the gaps found before the lender finds them.
Credit · In developmentApplying for finance and want the evidence pack complete and consistent first.
Why it exists
A seller says the business earns AUD 140,000. Behind that number sit bank statements, four BAS, a P&L and a schedule of add-backs. Checking them by hand is slow, and gaps are the easiest thing to miss.
The seller’s claim against provisional adjusted EBITDAAUD 30,000 · 21.4% of the claim
Revenue against bank deposits and against BAS salesNot calculated — the reasons are stated
Reserve account statementsJuly to December 2024, requested from the seller
What you give · what we test · what you get
Choose a module to see the records it reads, the checks it runs on them and what you receive at the end.
Results shown are Dealensics engine output on the Harbour Lane synthetic case. What Acquire reads, in full
Separation is in development: this is how it is being built, shown on its synthetic case.
Results shown are Dealensics engine output on the MORGAN-01 synthetic case. What Separation reads, in full
Credit is in development: this is its design, shown on a synthetic design case whose results are specified, not engine output.
Results shown are the ones the Credit design specifies for the NORTHSTAR-01 synthetic case, not engine output. What Credit reads, in full
Source traceability
Move the lens across the seller’s claim. Underneath: the figure the records support, the AUD 30,000 they don’t, the add-backs, the blocked comparison and the statements that were never supplied — each with its file and row.
Synthetic case. The seller claims seller proposed normalised earnings of AUD 140,000 (harbour-lane-seller-add-backs.csv, row 5). The supplied P&L shows EBITDA of AUD 110,000 (row 18). Add-backs included: AUD 0, so AUD 30,000, or 21.4% of the claim, is not reconciled. P&L revenue against BAS G1 sales: unable to assess. Reserve account statements for July to December 2024 were not supplied.
Every file, account and period, listed first.
What the records cover, drawn to dates.
Only like with like — or the reason it can’t be.
The exact record that would settle it.
What you receive
Every Dealensics review is built to end in the same kind of report. These pages come from synthetic cases with known answers: Acquire’s and Separation’s are engine output; Credit’s are the results its design specifies.
Harbour Lane Cafe Pty LtdFY2025Synthetic case
Key finding · F-01 · Not reconciled
The seller’s claim of AUD 140,000 (add-back schedule, row 5) exceeds provisional adjusted EBITDA of AUD 110,000 (P&L, row 18). Add-backs included: AUD 0.
View the evidenceMORGAN-01Synthetic householdNot legal advice
Coverage · savings account ••8804 · Not supplied
Statements stop from 1 November 2023 to 31 January 2024. The gap is dated and listed as a document to request — never filled in.
View the evidenceNorthstar Example Fabrication Pty LtdFY2026NORTHSTAR-01 · design case, not engine output
Coverage · account ••2020 · Evidence missing
90 days have no statement, and a reissued statement leaves another 92 in dispute. Annual revenue-to-bank checks are blocked until they are resolved; no difference is calculated.
View the reportSynthetic case Harbour Lane. Figures and questions are Dealensics engine output.
When information is missing
A missing month is easy to overlook, and it quietly changes every total that depends on it. Dealensics draws every gap, stops the comparisons that depend on it, and asks for the record that closes it.
Reserve account ••5678: no statements for July to December 2024.
P&L revenue against bank deposits is not calculated. The reason is stated instead of a number.
Bank statement coverage has gaps.
A month without statements is not a month without money. Nothing is estimated or filled in.
0.00No zero, no estimate, no average in its place
The request names the account and the dates, written for the person who holds it.
To the sellerPlease provide reserve-account statements for 2024-07-01 to 2024-12-31 so the FY2025 account coverage can be assessed.
Your accountant, lawyer or broker
Dealensics does the slow part — collecting, placing and checking — so your adviser can spend their time on judgement instead of on spreadsheets.
Starts from the findings register and the question pack — the earnings bridge, GST working and add-back support — instead of a stack of files.
Separation is being built so they can work from an organised, dated record of the household’s finances. Dealensics gives no legal advice.
Credit is being built so they can submit a complete, consistent pack, with the gaps closed before the lender asks. Dealensics makes no lending decision.
You make the decision. Dealensics never recommends what to do.
Why not a chatbot and a spreadsheet?
A chatbot can summarise what it’s given and a spreadsheet can add up what’s pasted in. Neither tells you where a figure came from, or what was never supplied.
Why trust it
Trust comes from being able to check the work. So every result shows how it was reached, and the parts that can’t be reached are said out loud.
Totals, differences and percentages come from deterministic, decimal-safe code. A language model never does the arithmetic.
Figures in a finding come with their file and row, so your adviser can check them in seconds.
A gap is shown and dated. It is never filled in, estimated or treated as zero.
Supported, partially supported, not reconciled, unable to assess, evidence missing. No scores, no verdicts.
Acquire early-access reviews are assisted: a person checks the documents, the extracted figures and each finding before a report is released.
Our synthetic cases have known correct results, and the engine is checked against them before changes ship.
What happens to your documents
Dealensics isn’t accepting customer documents yet: nothing is uploaded through this website, and every example here uses synthetic records. This is how a review is being built to handle them, with what is in place today.
Each module lists the records it reads, so you can leave out what isn’t needed, like tax file numbers.
Lists publishedFiles are held in an access-controlled workspace for that review only.
Tested with synthetic dataEach file is checked, then placed by entity, account and period.
Synthetic formats onlyCode does the arithmetic; a person checks each finding before a report is released.
Reviewers being appointedYou receive the report and its workpaper.
Delivery route being agreedDeletion removes the files, extracted data and reports from Dealensics’ storage, keeping only a content-free record that the review happened.
Built · retention periods pendingBefore any customer documents are accepted, the retention periods, the reviewers, the delivery route and the provider controls must be confirmed. Please don’t send financial records through the website’s forms. How the website handles enquiries
What Dealensics doesn’t do
Dealensics prepares the evidence and the questions. Decisions — and the advice of your accountant, lawyer or broker — stay with you.
Dealensics provides preliminary financial evidence analysis based on the documents supplied. It is not an audit, valuation, legal opinion, tax opinion or recommendation to buy or sell a business. Material matters should be reviewed with appropriately qualified advisers.
Questions
The questions people ask before they trust a review with their records.
Buyers of a business and their accountants or advisers (Acquire, in early access); people separating and their lawyers or mediators (Separation, in development); and brokers and business borrowers preparing a loan application (Credit, in development).
The records behind the numbers. For Acquire: the P&L, BAS, bank statements for every business account, the seller’s add-back schedule and its support. Each module page lists what it reads. Dealensics isn’t accepting customer documents yet, so please don’t send any.
None are being accepted yet, and nothing is uploaded through this website. Reviews are being built so that documents sit in an access-controlled workspace for that review only and can be deleted on request, leaving a content-free record. Retention periods, reviewers and provider controls must be confirmed before customer documents are accepted. Our security policy bars training Dealensics models on customer documents without explicit consent.
Whether the figures in front of you are supported by the records behind them: a seller’s earnings claim against the P&L, add-backs against their evidence, revenue against bank deposits and BAS — and whether the records cover the whole period at all.
A findings report — what was supplied and what it covers, each finding with its status, source and limits, and the questions to send next — plus a workpaper with the calculations.
Yes. The sample reports use synthetic cases with known answers. Harbour Lane (Acquire) and MORGAN-01 (Separation) were run through the Dealensics engine. NORTHSTAR-01 (Credit) is a design case: Credit is in development, so its results are the ones the Credit design specifies, not engine output.
A chatbot can summarise what it’s given, but it can’t show where each number came from or notice what’s missing. A spreadsheet is only as good as the formulas someone builds. Dealensics checks coverage first, compares only like with like, and keeps each figure next to its source.
They stay in charge. Dealensics prepares the evidence and the questions; your adviser applies their judgement; you make the decision.
It doesn’t value a business, score a deal or a borrower, recommend what to do, give legal, tax or financial advice, or make findings about any person. It isn’t an audit.
Each figure in a finding carries the file and row (or page) it came from, and each calculation is set out in the workpaper. Your adviser can open the source and check the number themselves, instead of trusting a summary.
The gap is dated and shown, comparisons that depend on it are blocked rather than guessed, and you get the exact request that would close it.
Calculations are done by code, each figure keeps its source, missing data is never treated as zero, and Acquire early-access reviews are assisted: a person checks each finding before a report is released.
To discuss an Acquire review, use the early-access enquiry. Separation and Credit are in development and not yet available. Not ready yet? Open the sample reports to see what a review produces.
What to do next
For buyers and their advisers. Early-access reviews are assisted: a person checks each finding before a report is released.
Discuss an acquisition review with Dealensics AcquireIn development and not yet available. Its page shows how it is being built, on a synthetic case.
Explore SeparationIn development and not yet available. Its page shows the design, on a synthetic design case.
Explore CreditSee exactly what a review produces first: three sample reports from synthetic cases.
Open the sample reports