Northwind is a 380-person construction-equipment dealer with offices in nine states. Their accounts-payable team of five spent the last week of every month doing the same thing: matching 1,200 vendor invoices against purchase orders and goods receipts in NetSuite, flagging the ones with variance, and chasing approvals.
It took three weeks. By the second week of May they were already behind on June.
We deployed Vera, an AI worker scoped tightly to AP reconciliation. Vera pulls invoices from the shared inbox, extracts line items, three-way matches each one against the PO and goods receipt in NetSuite, and routes the result.
Vera does not invent matches or guess. When she is not sure, she asks. Every action is logged with the human owner who approved it.
Median time to clear an invoice fell from 4 hours to 42 seconds. 87 percent of invoices clear without a human reviewing them. The remaining 13 percent are exactly the ones the AP team wants to look at: real variances, vendor disputes, missing paperwork.
The AP team has not shrunk. Two of the five have moved to vendor management work that was always on the roadmap and never got done. The month-end close is now under two business days.
Sara Patel, Northwind's CFO, told us the part she did not expect was the policy graph. Every exception Vera surfaced, every approver decision, every addendum cited, fed back into the workflow. By month three, last month's hard cases were this month's automatic clears. The agent got better at Northwind specifically, on Northwind data, without anyone retraining a model.
"Vera got better at Northwind specifically, on Northwind data, without anyone retraining a model."
If you want to see what a similar deployment would look like for your AP, AR, or close process, book a demo and we will scope it on the call.