Cut order-to-invoice time from three days to under an hour.
A 40-person freight brokerage ran orders through email, a legacy dispatch system, and accounting software that never spoke to each other. Three staff spent most of the day rekeying between them. We built a single order pipeline and connected it to both systems. Nobody lost a job — the same team now handles roughly a third more volume.
- Sector
- Logistics
- Engagement
- Custom software + integration
- Duration
- 14 weeks
- Practice
- Software Engineering
- Order to invoice, from 3 days
- 47 minOrder to invoice, from 3 days
- Manual steps removed
- 71%Manual steps removed
- More volume, same headcount
- 30%More volume, same headcount
01The situation
Orders arrived by email and phone, were entered into a dispatch system built in the early 2000s, and were then entered a second time into accounting to raise an invoice.
Three people spent most of their day on that transfer. Invoices went out an average of three days after delivery, which pushed payment terms out by the same amount.
The dispatch system could not be replaced — it held fifteen years of history and the operations team knew it well.
02What we did
We spent two weeks with the operations team documenting the real path an order takes, including the exceptions that never appear in a process diagram.
Rather than replace the dispatch system, we put a pipeline in front of it: orders are captured once, validated, and written to both dispatch and accounting automatically.
Exceptions route to a review queue instead of failing silently. Anything the system is not confident about is handled by a person, which was a condition of the operations team trusting it.
We migrated in stages by customer, so a problem would affect one account rather than the whole book.
03What changed
Invoices now go out under an hour after delivery. The finance team stopped chasing paperwork and started chasing payment.
The three staff who handled data entry moved to account management. Volume grew about 30% over the following two quarters without adding headcount.
The review queue averages under ten items a day, which is the number the operations team wanted to keep visibility of.
Built with
- TypeScript
- PostgreSQL
- Accounting API
Let’s build something that matters.
Thirty minutes. Tell us what’s slowing the business down and we’ll tell you what we’d do about it — whether or not you hire us.