Hundreds of hours a year won back on ZUS and PIT contributions

The challenge

Every month, the same routine. An accountant opened one program after another, read off each client's ZUS contribution and PIT figures, then manually retyped them into individual emails. With ~200 clients that added up to 24 hours a month - work that does not forgive mistakes.

The problem wasn't knowledge - it was retyping. Every manual transfer of an amount is a chance for a slip: transposed digits, the wrong client, an outdated rate. A single mistake like that can cost more than a full day of savings, because it has to be found, explained and fixed.

• 24 hours a month on a task that adds nothing beyond retyping numbers.
• A typo risk that grows with every additional client.
• Work piled into a few days each month - a bottleneck against ZUS and tax-office deadlines.

The solution

We built two things that work together: our Agregator składek ZUS+PIT desktop app and a workflow in n8n. The app pulls the contribution figures straight from the firm's accounting software - WAPRO Kaper and Płatnik. The workflow maps them to the right clients and prepares draft emails - each with the correct figure, addressed to the right person.

1. The accountant opens the Agregator and selects clients, or filters them by payment deadline (the 15th or 20th of the month).
2. The app pulls each client's ZUS and PIT-4 figures from WAPRO Kaper and Płatnik and sends them to n8n.
3. The workflow matches the client's tax ID (NIP) to an email address in a Google Sheet and builds a per-client email with the breakdown of figures.
4. The email is created as a draft in Gmail - the workflow doesn't send it. Every draft is logged, so the same company won't get two emails in the same cycle.
5. Finally, the firm gets a summary email: how many drafts were created and which clients are missing an address in the sheet.

The full monthly set of contributions for every client is generated with a single click. The figures are never retyped by hand, so the whole class of errors that came purely from retyping is gone. The workflow runs on mapped parameters, not on memory.