What counts as proof of income
Proof of income is any document that evidences what someone earns, and the acceptable list is longer than most people realise. The classics are recent pay stubs and bank statements showing income deposits; behind them sit tax returns and assessments for a full prior year, employer letters confirming salary, and — for those outside conventional employment — benefit or pension award letters, and accounts or invoices for the self-employed. Which combination is required depends on who's asking and why, but the documents themselves share one trait: they hold the numbers someone needs and offer no easy way to use them.
That's the problem this page is about. Collecting a folder of proof-of-income PDFs is only step one; the actual work is reading the figures out of them — the pay, the deposits, the annual totals — and that's where retyping and errors creep in. Converting each document into structured data turns a stack of evidence into numbers you can total, compare and check in minutes, whichever forms the applicant supplied.
| Document | What it proves | Best for |
|---|---|---|
| Pay stub / payslip | Employer pay, gross/net, YTD | Salaried employees |
| Bank statement | Income actually received | Anyone — proves money landed |
| Tax return / assessment | Full prior-year income | Self-employed, complex income |
| Employer letter | Stated salary / position | Salaried, as a supplement |
| Benefit / pension letter | Award amounts | Non-employment income |
| Accounts / invoices | Business income | Self-employed, contractors |
Why turn proof documents into data
A pile of PDFs proves income to a human who reads every page; it proves nothing to a spreadsheet or a system. The moment you need to total income across documents, compare a stub to a deposit, or feed a figure into a model, the documents have to become data. Doing that by hand is slow and risky — every number transcribed from a stub or a statement is a number that can be keyed wrong, and an income check built on a transcription error is worse than useless.
Converting the documents removes that step. Each pay stub becomes its earnings, deductions and totals; each statement becomes dated, signed transactions; each form becomes its labelled boxes. From there the figures are sortable, totalable and checkable, and the same conversion that captures them validates them. The result is that the proof you collected becomes proof you can actually compute on — which is the whole point of asking for it. The entry points are pay stub to Excel, the bank statement converter and financial statement to Excel.
Bank statements: proof the money arrived
Among proof documents, bank statements are uniquely powerful because they show income that genuinely landed, not just income that was promised. A pay stub says the employer paid; the statement proves the net pay hit the account, on the date, in the amount. For anyone whose income isn't a single salary — multiple jobs, benefits, side income — the statement is often the only document that captures all of it in one place.
Converting statements turns that proof into figures: recurring credits, total and average monthly income, and the exclusions (transfers, refunds) that keep the number honest. This is the heart of income verification from bank statements, and it's why statements are usually requested alongside or instead of stubs. Because the converter reads any bank into the same structure, an applicant's statements become analysable regardless of where they bank.
Pay stubs: the employer's figures
Pay stubs are the most direct proof for salaried employees, stating gross pay, each deduction and tax, net pay and year-to-date totals straight from the employer's payroll. Converted to data, a stub yields all of those as labelled fields, and a sequence of stubs annualises cleanly from the year-to-date column — useful when you need a yearly figure from monthly evidence.
FlowParse reads stubs from any payroll provider by meaning rather than a fixed template, so ADP, Gusto, Paychek, in-house and international formats all produce the same fields — see pay stub to Excel and, outside the US, payslip to Excel. The gross-minus-deductions-equals-net check runs automatically, so a misread stub is flagged rather than trusted. Stubs are strongest paired with statements: the stub's net pay should match a deposit, and that cross-check is one of the most reliable ways to confirm both.
Proof of income when you're self-employed
Self-employed and freelance applicants are where proof of income gets hard, because there's no payslip. Their evidence is some combination of bank statements showing client payments, a tax return or assessment confirming a prior year, and sometimes business accounts or invoices. Each is a document FlowParse can convert: statements into income credits, returns and accounts into their line items and totals.
The practical approach is to triangulate — read the average income from several months of statements, confirm it against the tax return or Self Assessment, and use accounts or invoices for detail where needed. Converting all of them into one consistent dataset is what makes a fair assessment of irregular, multi-source income possible without an analyst reading every page by hand. It's also why self-employed applicants are usually asked for more documents — and why turning them into data quickly matters most here.
| Self-employed document | What it adds | Convert with |
|---|---|---|
| Bank statements | Income actually received, month by month | Bank statement converter |
| Tax return / assessment | Confirmed full-year income | Financial statement to Excel |
| Business accounts (P&L) | Profit after costs | Profit-and-loss converter |
| Invoices | Detail of what was billed | Invoice PDF to Excel |
How to turn proof documents into data
Whatever the mix of documents, the workflow is the same. Gather everything the applicant supplied, convert each document with the matching tool, and you have one structured dataset to analyse — pay as fields, statements as transactions, forms as boxes. Because the same engine handles every type, a folder of stubs, statements and a tax form doesn't need three separate processes; it's one conversion pass.
From there you total, cross-check and compute whatever your process needs, with everything traceable back to the source document. The detailed method for the statement side is in the guide to verifying income; the principle across all proof documents is the same — convert first, then analyse data rather than reading PDFs.
Gather the documents
Collect every proof the applicant supplied — stubs, statements, forms; scans and photos are fine.
Convert each type
Run stubs, statements and forms through the matching converter for clean fields.
Cross-check the sources
Confirm a stub's net pay against a statement deposit; reconcile totals against a tax figure.
Compute and decide
Total income, apply your policy, and keep the structured data as your audit trail.
Accurate, validated figures
Proof of income only does its job if the figures are right, so validation is built in. Statements are balance-checked so a dropped or duplicated transaction is caught; pay stubs have the gross-minus-deductions-equals-net check; and extraction runs at around 98% field-level accuracy on standard documents with low-confidence fields flagged for review. Nothing is trusted blindly, and every figure traces back to the document it came from.
That traceability is what makes a converted income figure defensible. You can show exactly which deposits, which stub lines, or which form boxes produced a total — which matters when an applicant disputes a decision or an auditor reviews the file. Scanned and photographed documents are handled via OCR with confidence scoring, so even a phone photo of a stub becomes checkable data rather than something to retype.
Checking proof documents for consistency
Proof of income is a target for manipulation — an edited stub, a doctored statement, an inflated figure — and no converter can certify that a document is genuine. What FlowParse can do is surface internal inconsistencies that authentic documents don't have: a statement whose balance no longer reconciles, a stub where gross minus deductions doesn't equal net, or figures that contradict each other across documents. The balance and arithmetic checks in validation flag exactly these.
The strongest authenticity check, though, is cross-referencing the sources you already have. A stub's net pay should appear as a deposit; a year of statement income should be in the ballpark of the tax return; benefit letters should match benefit credits. When the documents are structured data, these comparisons are quick and the contradictions obvious. None of it is a fraud guarantee, but together it catches careless edits and tells you where to demand originals — with the judgement on authenticity remaining yours.
Who asks for proof of income
A lot of decisions hinge on proof of income. Lenders require it for mortgages, auto loans and personal credit; landlords and letting agents require it to confirm a tenant can afford the rent; employers verify it during onboarding; and benefit, grant and subsidy administrators check it against eligibility thresholds. In every case they receive a pile of documents and face the same chore — reading the figures out of them.
For these requesters, converting proof documents to data is the difference between an analyst reading every page and a fast, consistent check. The income lands in the same fields whatever the applicant supplied, so the assessment step doesn't change from one person to the next. The high-volume versions — a lender's origination flow, a large letting agency — benefit most, but even an occasional check is faster when the proof is structured data.
Preparing proof of income to supply
The other side of the transaction is the applicant — or the broker or accountant acting for them — who has to assemble convincing proof. Here converting documents helps too: a clean, totalled summary of income drawn from statements and stubs, presented alongside the originals, makes an application easier to assess and quicker to approve. For self-employed applicants especially, a tidy income picture built from converted statements can be the difference between a smooth application and a stalled one.
Brokers and advisers do this routinely on clients' behalf, turning a messy folder into a clear income story. The same converter that a lender uses to read proof, an applicant can use to present it — convert the statements and stubs, build the summary, and supply both the figures and the source documents. It's the same data either way; converting it just makes the proof speak for itself.
What this does — and what it doesn't
To be clear about the boundary: FlowParse converts and validates proof-of-income documents into clean, structured figures — pay, deposits, totals, with consistency checks. It is a data tool. It does not decide whether the proof is sufficient, does not approve or decline an application, does not score credit, and does not certify that a document is authentic beyond the internal checks it can run.
Those judgements belong to the party making the decision, because they depend on policy, risk appetite and rules a converter has no business encoding — and because accountability for the outcome rightly sits with whoever takes the risk. FlowParse's role is to produce the cleanest, most checkable input: accurate figures, validated and traceable, from whatever documents were supplied. That makes the decision better-informed and faster, while leaving the decision itself with you.
Handling sensitive income documents
Proof-of-income documents are deeply personal — pay, account details, benefit status — so handling them carefully is essential. Uploads run over TLS on EU-hosted infrastructure, the original document is deleted right after processing, files are isolated per user, and documents are never used to train AI models. You keep the structured figures; the source PDFs don't linger.
For organisations processing proof at volume, the document extraction API keeps applicant data within your own flow, with per-key authentication and usage logging for a clear audit trail. Handling other people's income evidence to a high standard is the baseline that makes using a converter for these documents appropriate in the first place.
Into your spreadsheet, model or file
Converted proof-of-income data goes wherever your decision is made. Export a clean Excel or CSV summary for review, push structured JSON into your origination or screening system over the API, or keep the figures in a spreadsheet alongside the originals as your audit file. One conversion, whichever destination your process needs.
Because the schema is consistent across document types and banks, a mixed pile of stubs, statements and forms produces one coherent dataset rather than three incompatible ones. Convert once, total and cross-check, and retain the structured rows for the record — that's proof of income turned from a folder of PDFs into evidence you can actually work with.
Turn proof of income into usable data
Convert pay stubs, bank statements and tax documents into clean, validated income figures you can total, cross-check and defend — from any provider or bank.