Converting a bank statement into Excel has become something anyone can do in about five minutes. If you can export a statement from your bank, you can get one row per transaction and columns you can sort, total, and send to a bookkeeper. That holds for most banks, whether the file is a PDF, a scan, or a download from online banking.
Some banks already offer CSV or Excel exports of your transactions. That is the fastest path when it exists, and the dates and amounts usually match the official record. The limitation is that those exports almost never include categories or expense types. You get raw rows with no “materials,” “fuel,” “software,” or “transfer.” If you need those labels to calculate and capture tax deductions, you still have to add them. Bank statement categorizers exist for exactly this reason.
If your bank will only give you a PDF, or you have a scan of an older statement, you can still get to Excel. The two practical options are a chatbot you already use (ChatGPT or Claude) and a dedicated bank statement converter.
How to use ChatGPT or Claude to convert a bank statement
Bookkeeping tasks like this are easy in ChatGPT or Claude now. Upload the PDF, use the prompt below, then copy the table into Excel. If you already pay for ChatGPT Plus or Claude Pro, you are not adding a subscription. There is almost no learning curve. You already talk to it. For a short monthly statement, it is fast.
The prompt is the part that matters. We have included a sample prompt below. Adjust it so it captures the exact fields you want. On a bank statement that is usually date, description, debit, credit, and balance, but banks do not all print the same columns. You may need to fine-tune the prompt for that statement.
Once the workflow looks right, ask ChatGPT or Claude to turn it into a skill. After that, you run the same prompt again by asking it to use the skill, and the process will run like clockwork.
A few limits. You are putting banking data on a public server. The output is not always consistent. Uploads are also capped, so a long file can miss rows or hit a limit. ChatGPT or Claude is enough for a single short statement you will check against the PDF by hand. In these cases, there are specific tools designed for bank statement conversion.
Sample ChatGPT prompt to convert a bank statement to Excel
Copy the prompt, attach your statement PDF, and send. Adjust the variables and categories as needed.
“Convert the attached bank statement into an Excel workbook.
Extract every transaction and create one row per transaction with these columns:
1. Transaction Date — YYYY-MM-DD
2. Posting Date — YYYY-MM-DD, if available
3. Vendor — cleaned merchant/payee name
4. Description — concise explanation based only on the statement
5. Original Description — exact transaction text from the statement
6. Amount — numeric value with no currency symbol
7. Currency — ISO code such as USD, EUR, or GBP
8. Transaction Type — Debit, Credit, Fee, Interest, or Transfer
9. Category
10. Subcategory
11. Confidence — High, Medium, or Low
12. Page Number
Use negative amounts for money leaving the account and positive amounts for money entering it.
Categorize transactions using these categories:
– Income
– Rent and utilities
– Office expenses
– Software and subscriptions
– Travel
– Meals and entertainment
– Transportation
– Advertising and marketing
– Professional services
– Insurance
– Taxes
– Bank fees
– Transfers
– Loan payments
– Refunds
– Personal
– Other
Requirements:
– Do not omit transactions.
– Do not include opening balances, closing balances, subtotals, or informational lines as transactions.
– Do not guess missing data. Leave the cell blank and use Low confidence.
– Preserve the original transaction wording.
– Normalize recognizable vendor names, but do not invent vendor details.
– Identify likely duplicates in a separate “Review” column.
– Ensure dates and amounts are stored as real Excel date and numeric values.
– Sort transactions chronologically.
– Add a second worksheet named “Summary” showing totals by category, currency, and month.
Before creating the workbook:
1. Report the number of transactions detected.
2. Reconcile the extracted transactions against the statement totals.
3. Flag illegible, ambiguous, duplicated, or unreconciled entries for review.
Return the completed file as an .xlsx workbook.”
When should I use a bank statement converter?
A general AI assistant can extract transactions from a clear statement, but you must define the columns, category rules, and formatting each time. Results may vary between documents, making monthly files harder to combine. If you run into consistency or output quality issues with ChatGPT or Claude, you should consider a bank statement converter.
A dedicated bank statement converter provides a repeatable workflow: upload one PDF or a batch, review the extracted transactions, and export them to CSV or Excel. It keeps fields such as dates, vendors, amounts, currencies, categories, running balances, and account periods in the same format every time. This consistency matters when combining months into one workbook, creating reports, or importing data into accounting software.
Dedicated tools are built to process quarters or years of statements, not just one page at a time. They can preserve more statement data, categorize expenses, identify uncertain or duplicate transactions, and help reconcile extracted totals against opening and closing balances.
For a one-off conversion, a carefully prompted AI assistant may be enough. When statement conversion becomes a monthly task or involves multiple accounts, a converter like ReceiptsAI can make economic sense because it reduces repetitive prompting, cleanup, formatting, and error checking.
A dedicated converter is the better habit if you do this every month, or if you need last year’s statements in one consistent file.
Conclusion
Getting a bank statement into Excel is no longer a special skill. Export from the bank if you can, use ChatGPT or Claude for a short PDF, or use a converter when you need the same columns every month with more accuracy and higher volume.