Why a private bank statement needs its own reading
A retail bank statement is, for extraction purposes, a fairly predictable document: a running list of dated transactions, a debit or credit column, a balance. A private bank statement rarely looks anything like that. It's often a relationship summary — a narrative overview from a relationship manager — sitting directly beside holdings detail and transaction activity, formatted however that specific institution chooses to present it, with no shared standard across banks.
A converter built for ordinary transaction ledgers tends to either miss the holdings entirely, misread a relationship-summary paragraph as a transaction row, or blend a mixed sub-account view into one figure that loses the underlying detail. This page describes how a private bank statement is read on its own terms — separating narrative from data, and holdings from transactions — before anything is exported.
That distinction — reading content by what it actually is, not by where it sits on the page — is what the rest of this page walks through in detail.
Where ordinary converters get it wrong
A relationship-manager paragraph read as a transaction
Narrative text with a dollar figure mentioned in passing — a portfolio value quoted in a summary sentence — gets misread as a row in the transaction table.
A blended sub-account figure treated as one balance
A checking sub-account and an investment sleeve reported together under one relationship total, with the split lost in a generic extraction.
Holdings dropped entirely
A converter tuned for a debit/credit ledger has nowhere to put a securities holdings table, so it's skipped rather than extracted.
A multi-currency figure silently converted
A holding priced in a currency other than the account's base currency gets flattened into one number, losing which currency it was actually reported in.
Why this isn't a simple text extraction
Pulling text off a page is the easy part — nearly any tool can do that. Deciding which text is a transaction, which is a holding, which is a balance, and which is narrative that shouldn't be treated as data at all is where a private bank statement diverges sharply from an ordinary bank statement to Excel conversion — the layout simply carries more kinds of content, mixed together, than a standard transaction ledger does.
What a private bank statement actually contains
| Section | What it typically contains |
|---|---|
| Relationship summary | A narrative overview, sometimes with a headline portfolio value — not itself a data table |
| Holdings detail | Individual positions with quantity, price and market value |
| Transaction activity | Deposits, withdrawals, trades and fees for the period |
| Sub-account breakdown | Checking, sweep and investment sub-accounts, sometimes shown separately, sometimes blended |
Each of these gets read for what it actually is, rather than forced into a single generic transaction table — narrative stays narrative, holdings become their own structured rows, and transactions are extracted as transactions.
What gets read
| Field | Source |
|---|---|
| Account holder, account identifiers | Statement header |
| Opening and closing balance | Summary section |
| Holdings and their values | Holdings detail |
| Deposits, withdrawals, trades, fees | Transaction activity |
How it works
Upload the private bank statement
Whatever mix of summary, holdings and transaction pages it contains.
Content is separated by kind
Narrative, holdings and transactions identified and handled distinctly.
Each field is extracted
With a confidence score, and currency captured alongside every value.
Ambiguity flagged, not guessed
A blended sub-account figure or an unclear total is marked for review.
Exported
Excel, CSV or JSON, with every figure traceable to its source.
A statement, converted
A private bank statement opens with a relationship-manager paragraph noting the portfolio's overall performance, followed by a holdings table and a transaction section for the month.
| Content type | Count |
|---|---|
| Narrative paragraphs excluded from data | 1 |
| Holdings extracted | 14 |
| Transactions extracted | 9 |
| Figures flagged for review | 1 |
The one flagged figure turns out to be a holding priced in a secondary currency without a clearly printed exchange rate on the page — correctly left for manual confirmation rather than converted with an assumed rate that might not match what the bank itself used.
Multi-currency accounts
Private bank accounts frequently hold assets in more than one currency — a base account currency alongside foreign-currency cash or securities. Each figure is read in whatever currency the statement itself reports it in, with that currency captured explicitly rather than silently converted to a single blended total that would obscure which currency each figure actually belongs to.
Manual vs. automatic
| Manual | Automatic |
|---|---|
| Statement retyped by hand, narrative skimmed and skipped | Narrative, holdings and transactions separated automatically |
| Multi-currency figures converted or blended inconsistently | Currency captured per figure, never silently converted |
| Holdings sometimes left out of the spreadsheet entirely | Holdings extracted as their own structured rows |
| Redone by hand for a differently formatted statement | Same method applies regardless of layout |
From one statement to a full custodian roster
Converting a single private bank statement is useful on its own. Most people using this aren't converting one statement, though — they're working through several private bank and brokerage statements as part of a broader consolidated wealth report. Reading each statement the same way, regardless of how many are involved, is what makes that broader consolidation possible in the first place.
Who uses this
Family offices and private wealth teams
Private bank statements read correctly as part of a broader consolidation.
Individuals managing their own private banking relationship
A clean Excel export without a manual retyping exercise each period.
Outside accountants and bookkeepers
A structured starting point instead of a mixed-content PDF to work through.
Wealth advisors preparing client reviews
Holdings and transaction detail ready for analysis without a manual pull first.
Edge cases worth knowing
A statement that reports a holding's value as of a slightly different date than the rest of the account — common for less liquid positions priced less frequently — has that date captured alongside the figure, rather than presented as though every holding were valued on the same day.
A relationship summary that quotes a headline portfolio figure inclusive of assets held elsewhere — outside the specific account the statement covers — is recognized as commentary rather than treated as the account's own balance, since conflating the two would overstate what that specific statement actually reports.
Why a traceable figure beats a rounded one
An export that shows only final numbers, without the statement page each figure came from, is easy to trust until someone asks why a specific total looks the way it does. Keeping the source page and a confidence score attached to every figure turns that question from a scramble back through the PDF into a detail already sitting in the data.
PDF, scan or photo — the same reading
A private bank statement doesn't always arrive as a clean, digitally generated PDF. Some relationships still mail paper statements that get scanned by an assistant; some are forwarded as a phone photo when a quick look is all that's needed at the time. Each of these goes through the same reading process — OCR first for anything that isn't already digital text, then the same field extraction and separation of narrative from data — with the same confidence scoring applied regardless of source quality.
A blurry scan or an angled photo doesn't silently produce a confident wrong answer — it produces a lower confidence score on the affected fields, flagged for a second look, which is a meaningfully different (and safer) failure mode than a clean-looking export that quietly got a figure wrong.
Handling the relationship-manager narrative
Many private bank statements open with a page or two of prose from the relationship manager — market commentary, a note on recent portfolio activity, sometimes a headline performance figure quoted in passing. None of this is account data, and none of it should end up in an export alongside real balances and transactions.
This narrative content is recognized as commentary and excluded from the structured output entirely, rather than either being dropped silently (which would hide that it existed) or misread as data (which would corrupt the real figures). The distinction matters most on statements where a dollar figure appears in the narrative itself — a quoted year-to-date return, for instance — since that figure describes performance commentary, not a line item the export should treat as a balance or a transaction.
What this doesn't do
Doesn't calculate performance or returns
It extracts balances, holdings and transactions — performance calculation is a separate step.
Doesn't apply your own currency conversion
It captures each figure's own reported currency rather than converting to a single base currency.
Doesn't give investment or tax advice
Any question about allocation, performance or tax treatment belongs with your advisors.
Doesn't replace the private bank's own statement as the record of account
It's a structured working copy — the bank's own statement remains the authoritative record.
What the exported workbook looks like
The export separates what a private bank statement actually contains into distinct, clearly labeled sheets rather than forcing everything into one flat table: a summary sheet with account identifiers, opening and closing balances and currency; a holdings sheet with each position, its quantity, price and market value where the statement provides them; and a transactions sheet with every deposit, withdrawal, trade and fee for the period.
Every row on every sheet carries the confidence score and source page reference described above, visible directly in the workbook rather than tucked away in a separate log — so a figure worth double-checking is easy to spot without cross-referencing a second document.
Where this fits in a broader wealth report
Converting one private bank statement solves an immediate problem — a document that resists a simple copy-paste — but most people arriving at this page have a bigger task just past it: rolling that converted statement together with a handful of others into one net worth or allocation view. That broader step, and the internal-transfer matching it requires, is covered in the consolidation guide, using the same field-level reading described on this page as its foundation.
Treating conversion and consolidation as two connected steps rather than one — read each statement correctly first, then line the results up together — is what keeps a family's wealth report accurate even as the number of custodians involved grows well past what a single statement export covers on its own.
Security and privacy
Uploads are encrypted with TLS from end to end.
Processing runs on infrastructure with SOC 2-aligned controls.
Original documents are deleted shortly after processing.
Nothing you upload is ever used to train AI models.
For statements that reveal the full shape of a private banking relationship, that matters — details are on the security page.
Checking the first conversion against what you already know
Before relying on this for a statement you haven't reviewed carefully by hand, run it once against a statement you already know well — a recent one where you can confirm every balance and holding from memory or an existing record. That first check builds confidence in the specific way your bank's statements are read, and surfaces quickly if any part of that particular layout — an unusual sub-account grouping, a footnote convention — needs a closer look before you trust the output on a period you haven't already reviewed yourself.
The short version
A private bank statement is one of the harder financial documents to convert well, precisely because it mixes narrative, holdings and transactions in a way an ordinary bank statement never does. Reading each of those three content types for what it actually is — rather than forcing everything into one flat transaction table — is what separates a genuinely useful export from one that technically pulled some text off a page.