Auto-Categorize Your Expenses With AI (Right in Your Spreadsheet)

Export your bank or card transactions, paste the descriptions into Claude or ChatGPT alongside your own list of categories, and the AI labels every row in one pass. An hour of squinting at cryptic merchant names becomes a single review.

Month-end expense sorting is nobody's favorite hour. Hand your transaction list and a fixed set of categories to any capable AI, and it sorts every row in one pass while you just review.

Your bank export is full of lines like SQ *BLUE BOTTLE or AMZN MKTP US*2X4. A human has to decode each one. An AI already knows SQ * is a Square merchant and AMZN MKTP is an Amazon purchase. Feed it the description and it assigns a category almost instantly. The trick that makes this reliable: hand it your category list instead of asking it to invent one.

Before you touch any AI, decide your categories. A fixed list is the single difference between tidy books and forty slightly different labels. Keep it short and concrete:

• Software, Travel, Meals, Office
• Marketing, Contractors, Bank Fees, Other

Now every transaction has exactly one home. When the AI is unsure, it drops the item in Other and flags it, instead of guessing a brand-new label you have to clean up later.

Export your transactions from your bank or card as a CSV and open it in Google Sheets or Excel. Keep just three columns: Date, Description, Amount, then add an empty Category column. Now delete anything sensitive. You do not need full card numbers to categorize, so removing that column means you never paste it anywhere. Clean inputs make every step after this faster.

Copy your Description column and paste it into Claude or ChatGPT with this brief. The key instruction is "same row order", so the answers line up perfectly when you paste them back.

Categorize each transaction into exactly one of these:
Software, Travel, Meals, Office, Marketing, Contractors, Bank Fees, Other.
Return the category only, one per line, in the same row order.
If unsure, use "Other (check)" so I can review it.

Descriptions:
{paste your Description column here}

Drop the results into your Category column. Dozens of rows, sorted in seconds.

Now filter your sheet for anything marked (check) or Other. These are the genuinely ambiguous merchants, a name that could be a client lunch or groceries. Fix just those by hand. This focused review takes a couple of minutes versus the half hour it takes to label every row from zero. You spend your attention only where judgment is actually needed.

The payoff is the report. Ask the AI: "From this categorized list, total the Amount by category and tell me my three biggest spending areas this month." In seconds you get a clean breakdown, and your sheet is ready for your accountant. While you are there, ask it to list recurring charges, that is how you catch the subscription you forgot you were paying for.

Open last month's statement and export it as a CSV. Trim it to Date, Description, Amount, then run the prompt above with your own category list. Reviewing one real month is the fastest way to see how much bookkeeping this quietly removes, and to fine-tune your categories before next month.

Paste twenty real transactions into a sheet and run the categoriser on those before wiring anything permanent. Twenty is enough to surface every failure mode your bank's naming will throw at it.