We automated seller settlement for influencer group-buy vendor Wooha Company. Registering seller info → tallying sales quantities from settlement statements → dispatching each seller's statement — work that took hours by hand for hundreds of sellers now takes seconds to minutes.
As suppliers and sellers grew, so did the settlement team's repetitive manual work.
Each supplier and platform used a different file format, and even the same product had different option labels — so the owner had to cross-check by eye "which seller's which option is this" and transcribe quantities one by one.
Building a statement per seller, capturing or copying it, and sending it one at a time over KakaoTalk. With dozens or hundreds of sellers, sending alone took hours.
Entering names, accounts, and business numbers directly into spreadsheet cells made it easy to put values in the wrong cell or overwrite existing ones.
The exact order the staff already followed — register sellers → process statements → dispatch statements — turned into uploading a file and pressing a button on the web.
Fill in the form and it's recorded accurately into the master sheet.
Upload files and quantities are keyed in per seller and option.
Each seller's statement is auto-generated and sent in bulk.
Enter a new seller's name, account, and business number in the web form and press "Register," and it's recorded to the exact position in the master group-buy sheet. No opening the spreadsheet to hunt for a cell.
Upload the settlement files (Excel · PDF) from suppliers all at once, press "Start processing," and it reads each file for per-seller, per-option sales quantities and fills the right cells in the master sheet automatically.
Pick the settlement month and press the button, and each seller's statement is auto-generated in the same format, saved to the monthly Drive folder, and emailed to every seller at once.
The goal was to automate repetitive work like settlement, so a small team can handle more sellers and revenue.
From the settlement statement suppliers send in (input) to the statement that goes out to sellers (output), here's a real processing example. (Sensitive info like accounts and resident numbers is masked; based on sample / dev data.)




Repetitive entry, cross-checking, and individual dispatch are gone; the staff just review. (Figures based on Wooha Company's own materials, 100-seller example.)
The 2–3 hours spent cross-checking and transcribing 100 sellers' settlement statements by eye now finishes in about 30 seconds — just by uploading the files.
We're jointly envisioning an expansion into an all-in-one "domain expert + AI agent" platform — seller sourcing monitoring, a customer-support (CS) chatbot, automated content (shorts) generation, and more.
The keys are document recognition that reads differently-formatted statements, and matching that finds the same seller and option despite different labels.