The workflow, in plain English
Before: Log into bank. Export CSV. Open Google Sheets. Import. Clean up formatting. Categorize. 20 minutes for 30 transactions.
After: Dassi reads transactions from your bank page. Switches to your Google Sheets tab. Finds the next empty row. Types each transaction: date, description, category, amount. 30 transactions in 20 seconds. No CSV. No import. No cleanup.
Why traditional tools can’t: Banks don’t have Zapier integrations. CSV export loses context. Dassi reads the live DOM and writes to the live spreadsheet — two web pages talking through an AI intermediary.
Step-by-step
- Open your bank transactions page in one tab.
- Open your Google Sheets budget in another tab.
- Ask dassi to transfer transactions into the next empty rows.
- Review the new rows and correct any category edge cases.
Example prompts to try
- “Copy the last 30 transactions into my budget sheet.”
- “Use my category list and flag anything uncategorized.”
- “Add a note column with merchant normalization.”
Tips for better results
- Be specific about what you want injected into the page (buttons, filters, a panel, a summary, etc.).
- If the page has multiple sections, tell dassi exactly what to focus on (e.g., “the transactions table” or “the diff for file X”).
- Prefer safe workflows first: draft, summarize, label, and prepare — then take actions (submit, purchase) only after review.