Microsoft 365 Copilot can talk about a bank statement, but the feature that actually pulls transactions out of the PDF is Power Query, and Microsoft documents exactly where it struggles: multi-line rows, scanned files, and large documents that time out. Upload the statement here and download a sheet whose closing balance foots.
Last updated September 2026
Upload your bank statement
Drop file here or click to upload
PDF, JPG, PNG, BMP, HEIC, TIFF, MT940
Uploading...
Microsoft 365 Copilot does not import a bank statement PDF into Excel. The feature that does is the Power Query PDF connector, at Data, Get Data, From File, From PDF, which Microsoft ships in Excel for Windows with Import as its only listed capability. It has no OCR, so a scanned statement returns an empty query, and Microsoft documents multi-line rows as a known limitation. Bank statements are largely multi-line rows. BankXLSX extracts every transaction in order, reads scans, and exports Excel, CSV, QBO, OFX, QFX or QIF.
None of this is a knock on Copilot, which is very good at the job it is built for. The gap is that people buy a Copilot seat expecting document extraction, and extraction is handled by a different, older feature with limitations Microsoft publishes plainly.
Copilot in Excel operates on data already in the worksheet. Ask it to import a statement and the real answer is Data, Get Data, From File, From PDF, which is the Power Query connector and has nothing to do with your Copilot license.
Microsoft connector documentation carries a Handling multi-line rows section and suggests cleaning up with UI operations or custom M code, naming Table.FillDown and Table.Group. On a bank statement those wrapped lines are wires, ACH addenda and long merchant strings, so the exception is most of the file.
The connector lists Import as its supported capability and no prerequisites. There is no OCR anywhere in it. A mailed statement or a phone photo has no text layer, so the query loads nothing and Excel gives you no warning about why.
For large PDFs Microsoft suggests editing the Pdf.Tables call to step through with StartPage and EndPage, and disabling MultiPageTables to avoid huge intermediate values. Loading a year of statements at once means moving to the Folder connector and a combine step.
The PDF connector and the Pdf.Tables function behind it are Windows Excel features, so Get Data on macOS simply does not list From PDF. Half a finance team can be locked out of the workflow by the laptop they were issued.
One job, held to the standard bookkeeping enforces: every row, in order, with the arithmetic intact and the same result every time.
Upload the PDF and get a finished sheet. There is no Navigator to guess your way through and no M code to write when a row wraps.
A wire with three lines of remittance detail comes back as a single transaction with its date and amount attached, not stranded across four rows.
Mailed statements and phone photos are read directly, so the files that return an empty Power Query still convert.
Opening balance plus credits minus debits is checked against the closing balance printed on the statement, so you can prove the extraction was complete.
Multi-file backlogs convert together instead of needing a folder connector and a combine step per account.
It runs in the browser, so the missing From PDF menu on macOS stops being a problem.
No connector, no Navigator, no credit card.
Drop in a PDF or scanned bank statement. Multi page and multi month files are fine.
Tip: Password-protected PDFs are supported.
Every transaction is pulled into structured columns automatically, balance included.
Tip: Scans are read with built-in OCR.
Compare the last balance in the sheet against the closing balance on the PDF. If it ties, the extraction is complete.
Tip: Run this check on any tool, this one included.
An honest side by side, including the work Copilot does better.
Clear a client backlog without hand-fixing wrapped rows in Power Query.
Produce workpapers that foot and can be reproduced next season.
Keep the analysis in Copilot and stop losing afternoons to the import step.
Convert statements without the From PDF menu Excel for Mac never had.
No, not on its own. Copilot in Excel works on data that is already in the worksheet, so it has nothing to act on while the transactions are still locked in a PDF. The Microsoft feature that opens the PDF is the Power Query PDF connector, under Data, Get Data, From File, From PDF, and it is a separate thing from Copilot. Once rows are in the grid, Copilot is genuinely good at the next part: categorizing, writing the formula, building the pivot.
| Microsoft feature | Opens a statement PDF? | Reads a scanned statement? | What you get |
|---|---|---|---|
| Copilot in Excel | No | No | Works on data already in the grid. It writes formulas, sorts, and charts what is there. It cannot fetch rows out of a PDF for you. |
| Copilot Chat with a file attached | Yes | Sometimes | Prose, a summary, or a partial table in the chat window. Useful for questions, not for a register a client will see. |
| Power Query PDF connector (Data, Get Data, From File, From PDF) | Yes, Windows only | No | Tables detected in the PDF, loaded into the sheet. Microsoft lists the supported capability as Import, with no OCR step. |
| BankXLSX | Yes | Yes, OCR built in | Every transaction in original order with a running balance column, as Excel, CSV, QBO, OFX, QFX or QIF. |
It can discuss one. Attach a statement to Copilot Chat and it will summarize spending, flag a large payment, or answer a question about the file. What it will not reliably do is reproduce every line in order with the balance intact, and that is the job bookkeeping actually enforces. Treat Copilot as the analyst and use a converter for the extraction, in the same order you would with ChatGPT or Claude.
On Excel for Windows, open a blank workbook, then Data, Get Data, From File, From PDF, pick the statement, and choose a table in the Navigator. Microsoft lists the connector under Excel with Import as the supported capability and no prerequisites. On a clean, text-based statement from a large bank this works. On the rest of them it needs a cleanup pass, and the reasons are specific rather than mysterious.
| What is on the statement | What Power Query does with it | Microsoft documented fix |
|---|---|---|
| A wire or ACH line whose description wraps onto two or three lines | Splits one transaction across several rows, leaving the date and amount stranded on the first one | Microsoft documents this as Handling multi-line rows and points to UI cleanup or custom M code, naming Table.FillDown and Table.Group |
| A statement your bank mailed, or a phone photo of one | Returns an empty query. There is no text layer to read and the connector has no OCR | None inside Power Query. You have to OCR the file elsewhere first, then import the result |
| A long statement, or one continuous register running over many pages | Slows down and can time out while it builds large intermediate values | Microsoft suggests editing the Pdf.Tables call to walk the file with StartPage and EndPage a page or a small range at a time, and turning MultiPageTables off |
| Twelve monthly PDFs for one account | The PDF connector takes one file at a time | Microsoft directs you to a multi-file connector such as the Folder or SharePoint Folder connector, then a combine step |
| You are working on a Mac | There is no From PDF entry under Get Data | Convert the file outside Excel, or shape it in Excel for Windows and open the workbook on the Mac |
Almost always because the PDF is an image rather than text. Open the file and try to select a transaction description with your cursor. If you can highlight the words, there is a text layer and Power Query has something to parse. If the whole page selects as one block, it is a scan, the connector finds no text, and the query comes back empty. The connector imports; it does not run OCR, so an image statement has to be read by something that does before Excel can use it.
If you are on a Mac, it is not missing, it was never there. The PDF connector and the underlying Pdf.Tables function are a Windows Excel feature, which is why Get Data on macOS shows no From PDF entry. On Windows, the option lives under Get Data, From File and needs a current Microsoft 365 build. There is a Mac route that does not involve Excel at all, and for a backlog of files a batch conversion beats repeating a manual import twelve times.
Power Query is included with Excel on Windows, so the import path costs nothing extra. Copilot is the paid part: Microsoft sells Microsoft 365 Copilot as an add-on that requires a qualifying Microsoft 365 base license, so a seat is never the only line on the bill. Worth knowing before you buy a Copilot seat expecting it to solve statement conversion, because that is the one part of this workflow Copilot does not do.
Not by itself. Copilot in Excel acts on data already in the worksheet, so it cannot reach into a PDF. The import is done by the Power Query PDF connector under Data, Get Data, From File, From PDF. Copilot is useful afterward, once the transactions are in the grid and you want them categorized or summarized.
Yes, in the sense of answering questions about one. Attach the statement to Copilot Chat and it will summarize spending or flag a payment. It will not reliably reproduce every row in order with a running balance, so it is a poor fit for anything a client, an auditor or a lender is going to check.
Two common reasons. On macOS the PDF connector does not exist, because Pdf.Tables is a Windows Excel feature. On Windows it sits under Get Data, From File, From PDF and needs a current Microsoft 365 build, so an older perpetual Office release will not show it either.
The PDF is almost certainly a scan. Try selecting a transaction description with your cursor: if the whole page highlights as one block there is no text layer, and the connector has no OCR to fall back on. Microsoft lists Import as the connector capability, nothing more, so image statements need OCR before Excel can read them.
That is the multi-line row limitation Microsoft documents on the connector page. Descriptions that wrap, like wire remittance detail or ACH addenda, get broken into separate rows with the date and amount left on the first one. The documented fix is manual cleanup or custom M code using Table.FillDown or Table.Group.
The PDF connector handles one file per query. Microsoft points you to a multi-file connector such as the Folder or SharePoint Folder connector, then a combine step to stack the results. It works, but it is a build, and it breaks again whenever a bank changes its statement layout.
It can time out. Microsoft advice for large PDFs is to edit the Pdf.Tables call and walk the document with the StartPage and EndPage options one page or small range at a time, and to disable MultiPageTables when the document is one huge table, because that option collects very large intermediate values.
No. Power Query ships with Excel on Windows at no extra cost, and BankXLSX runs in a browser regardless of your Microsoft licensing. Microsoft sells Microsoft 365 Copilot as an add-on requiring a qualifying base license, so buying a seat purely to convert statements pays for a capability Copilot does not provide.
The same tradeoff, measured on ChatGPT.
Long documents, the page cap, and dropped rows.
Google Gemini and the Drive workflow, compared.
Convert statements without the Windows-only connector.
OCR reads scanned and image statements.
Cancel anytime from your account settings · refund policy
Get started converting bank statements to spreadsheets.
USD
per month
per month
$288 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 2,500 pages |
| Pro AI Best accuracy | 500 pages |
Scale statement conversion across your team with automation.
USD
per month
per month
$888 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 10,000 pages |
| Pro AI Best accuracy | 2,000 pages |
Enterprise-grade bank statement conversion and controls.
USD
per month
per month
$ charged today
Choose speed vs accuracy when extracting
| Base AI Faster | pages |
| Pro AI Best accuracy | pages |
