Agents In Use
What’s new todayLast update
108 use cases added today

Use Claude to categorize bank statements in Notion

Raygar He runs two small businesses as a sole trader and wanted to see income, expenses and profit per project. Separate accounting products had failed to become a habit; Notion was already where he kept projects, clients and notes.

Source published raygarhe.comAgent: Claude
OUR ADAPTATION

Suggested prompt

Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.

Build a Notion income-and-expense tracker with my chosen categories, business labels and project fields. Write scripts to import quarterly bank statements and produce an accountant export.

Evidence at a glance

Sourced account
Source author
Raygar He
Original account
Read the source ↗
Published
Source link checked
Reported result
Working Notion views for client income, expense categories, project profit and financial-year earnings, with an accountant export and manual quarterly PDF inputs.
Human involvement
He chose categories, manually downloaded statements and sent exported views to his accountant.

A linked account supports attribution. It does not mean we independently reproduced its result.

Report an issue with this caseOpen correction email

Opens your email app with the case link included. Nothing is sent automatically.

How the agent did it

  1. Choose categories and database fields for both businesses.
  2. Have Claude create structure, transaction-classification and update scripts.
  3. Download quarterly statements and export the accountant's required view.

Raygar He runs two small businesses as a sole trader and wanted to see income, expenses and profit per project. Separate accounting products had failed to become a habit; Notion was already where he kept projects, clients and notes.

He described a database with date, amount, category, business and tax-deductibility fields. Claude Code wrote three scripts: one created the Notion database structure, another sorted bank-statement transactions into categories, and a third updated the records when new statements arrived. Software subscriptions, transport and client payments were examples of the classification work.

The setup took an afternoon, according to He, with most of his effort spent deciding categories. He retained a deliberately simple input routine. Invoices were labeled in Gmail, statements were downloaded quarterly as PDFs, and sensitive project files stayed on his machine. He considered a bank connection but decided four annual downloads did not justify its complexity or account access.

The resulting Notion views show income by client, expenses by category, project profit and earnings by financial year. When the accountant needs figures, He exports a view and sends it over. This is a working recordkeeping arrangement with manual collection and category decisions; the article does not report an independently reviewed classification accuracy rate.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: www.raygarhe.com · Raygar He

Frequently asked questions

Why use Notion instead of another accounting app?

He already opened Notion daily; forgetting to open a separate accounting product was the recurring problem.

Did it connect directly to the bank?

No. He retained quarterly PDF downloads, deciding that four downloads did not justify a bank API connection.

What did Claude actually build?

Three scripts for database structure, transaction categorization and updates from new statements.

More details

About this account

Claude wrote the database, classification and update scripts for He's existing Notion workspace.

Firsthand source opened and read. Suggested prompt written by us from the reported workflow; no original task prompt is quoted.

Credit: Raygar He

Source published

Added to Agents In Use

Source checked

Tools and topics

Claude Code · Notion · Bank PDFs · Gmail

Bank statements · Notion accounting · Project profit

Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.