AI operations for industrial, logistics, and manufacturing teams
Case study: Sales order automation for a steel manufacturer
Incoming purchase orders turned into ERP-ready sales orders, instead of retyped by hand.
01The situation
What order entry looked like
A steel manufacturer was taking purchase orders in whatever format customers sent: PDFs, spreadsheets, email bodies, and the occasional phone call. Sales ops retyped each one into the ERP.
The pressure was inconsistency, not volume.
- One customer sent clean PDFs. Another sent Excel with tabs everywhere.
- Grade, size, quantity, and delivery had to match what the mill could actually run.
- A mistyped line became the wrong order on the floor.
- Hours of cleanup sat in front of every order that hit the ERP.
They needed the orders structured and checked before they entered the system.
02What we built
Sales orders the ERP can take
The work was order intake, not a new ERP.
- 01Pull POs inEmail, PDF, spreadsheet, and the other formats customers already send.
- 02Extract the orderGrade, size, quantity, delivery, and the rest of the line, structured.
- 03Review against the originalSide by side with the document, so the team can catch what extraction missed.
- 04Flag errors before the ERPMismatched totals and incomplete lines, fixed before they become mill orders.
- 05Send to the ERPA clean sales order into the systems they already run.
03Results
What the work made possible
- 01Hours to minutesOrder entry no longer sits in a retyping queue.
- 02Errors caught firstMismatched totals flagged before they hit the ERP.
- 03Ops on the exceptionsThe team spends the time on the lines that need a person, not on keying every order.
Let’s talk
Schedule a free consultation →Made in America
Anvil Labs LLC
AI operations for industrial, logistics, and manufacturing teams.
anvil.industries