Vic.ai
78% of invoices, no human input
Countsy sells outsourced accounting to venture-backed startups. The firm handles accounts payable for hundreds of clients. Before the change, staff coded each invoice by hand at about five minutes per invoice, up to 3,500 invoices a month. Countsy adopted Vic.ai early as a pilot partner. The tool reads each invoice, predicts the ledger coding (cost account, department, location), and runs the approval flow inside the firm's NetSuite, Bill.com, and Expensify stack. The accountants moved from data entry to review.
How it wentThe vendor case study reports 78% of invoices now process with no human input. Processing time per invoice fell 84%, from 4 to 5 minutes to under 2 minutes. Coding accuracy reached 95%. The firm states no reliable speed data from before the change, and the results rest on one standard tool stack across all clients.
Try it · based on the source's workflow
Read each invoice in [folder or system]. Predict the coding fields: [cost account, department, location]. If your confidence is high, send the invoice to the approval flow in [accounting system]. If anything is unclear, flag it for my review instead of guessing. Show me a daily list of what you coded, what you flagged, and any invoice whose total changed.
A starting template based on the workflow the source describes. Review it before you use it; you are responsible for what it does.
78% of invoices, no human input — from The SmallOffice Brief, a daily brief of real AI cases for small businesses: https://smalloffice.space/case-78-of-invoices-no-human-input