Wilson by Cartage
Try Wilson →
AI logistics agent vs. traditional TMS

Legacy ERP freight automation: keep the ERP, fix the hand-off

Key Takeaways

  • Wilson, by Cartage, is an AI logistics agent that works from the order data a legacy ERP already produces.
  • Freight automation does not require replacing the ERP first.
  • An API is not required. A scheduled file, a spreadsheet, or email can carry the orders.
  • The feed needs an owner who notices when a file is late, incomplete, or sent twice.
  • A corrected order must update the existing shipment, not create a second one.
  • A scheduled file updates on a schedule. Plan the carrier work around that timing.
  • Returning updates to the ERP is a separate decision. It depends on scope.

Can freight automation work without replacing a legacy ERP?

Yes, freight automation can run on a legacy ERP as long as the orders reach it reliably. The ERP stays the system of record for orders. The freight work moves to a system built to do it.

Many companies run an ERP that is fifteen or twenty years old. It works. Finance trusts it. The plant runs on it. But it was never built to email carriers, compare quotes, or chase a proof of delivery. So the team does that work by hand, around the ERP.

Replacing the ERP to fix freight is a large bet. Core system projects often run long. In ARC Advisory Group research reported by Inbound Logistics in 2017, only 27% of TMS implementations finished on time with full functionality. Another 9% finished late and did not work as intended. Those figures cover transportation software, not ERPs. They still show the risk of solving a coordination problem with a system replacement.

The better first move is smaller. Find the order data the ERP already sends out. Make that feed reliable. Then hand the carrier work to an AI agent. For a wider view of what has to connect first, see Legacy ERP and modern freight: what has to connect first.

Four hand-offs between a legacy ERP and the carrier

Four hand-offs connect a legacy ERP to a delivered, paid load. Each one can fail in its own way. Each one needs an owner.

Hand-off Direction What often goes wrong Who owns the fix
Order released ERP to freight workflow Held orders slip into the export Operations
Load booked Freight workflow to carrier Weight or units differ from the packed shipment Logistics
Status and proof of delivery Carrier to freight workflow Updates arrive by email and never reach the order Logistics
Invoice and payment Carrier to finance The invoice cannot be tied back to the order Finance

The first hand-off is the one most teams worry about. It is also the one legacy systems handle best. Most old ERPs can produce a report or file of released orders. The weak spot is usually ownership. Nobody checks that the file arrived.

The last three hand-offs are where the hours go today. Carrier emails, status checks, and invoice matching all happen outside the ERP. An AI agent takes on that work. The ERP does not need to change for it to happen.

Feed health: four checks for a scheduled order file

A scheduled order file is reliable when it passes four checks every day. These checks keep a simple feed from becoming a hidden source of errors.

Arrival. The file shows up on time. If it does not, someone hears about it within the hour, not at the end of the day.

Completeness. Every released order is in the file. Every order has a ship-to, a ready date, a weight, and a handling-unit count.

Uniqueness. Each order appears once. A renamed file or a re-run export must not make an old order look new.

Corrections. A changed order arrives with the same order number. The freight workflow updates the existing shipment instead of starting another.

Timing is the fifth thing to settle. A file that runs at 6 a.m. and 2 p.m. updates twice a day. That is fine for most outbound freight. It is not instant, and nobody should describe it as instant. Orders released at 3 p.m. may wait for the next run. Plan carrier cutoffs around that.

Write down the owner of each check. A short line works: "Operations checks the 6 a.m. file by 6:30." That one sentence prevents most silent failures.

When the legacy feed is the wrong place to start

A legacy feed is the right start in most cases, but four situations call for fixing the data first. Use these rules before the first load moves.

  • The export includes orders on hold. Fix the filter first. Otherwise carriers get loads that cannot ship.
  • Weights are missing on most orders. Find where weights live today. If they live on the packing slip, bring that source into the feed.
  • The same order can appear under two numbers. Settle the order identity first. Two numbers for one order will create two shipments.
  • Nobody can say who runs the export. Name an owner first. An unowned feed fails quietly.

Outside these cases, start. A spreadsheet with consistent columns is a valid feed. So is an email with a standard attachment. The aim is a dependable hand-off, not an elegant one. For more on adding automation without a system swap, see Logistics automation without system replacement.

Example: a missed file and a corrected order on one Tuesday

One ordinary Tuesday shows how a scheduled feed should behave when two things go wrong. This is an illustrative example, not a customer case.

At 6:00 a.m., the ERP is due to drop the morning file. It does not arrive. A server restart stopped the export job. At 6:30, the operations lead notices the gap during the arrival check. IT re-runs the export. The file lands at 7:10 with 14 orders.

The freight work starts from that file. Quote requests go to approved carriers for all 14 loads. Twelve book by late morning. Two need a decision because no quote fits the rate target.

At 1:00 p.m., the plant finds that one booked order is short a pallet. The order is corrected in the ERP. The 2 p.m. file carries it with the same order number and a new weight. The freight workflow matches it to the existing shipment. It does not create a new one. Because the carrier already accepted the load, the change goes to a person. The person approves the revised weight with the carrier.

By the end of the day, three things are true. Every order became exactly one shipment. The missed file was caught within 30 minutes. The weight change has a recorded approver. That is what a healthy legacy feed looks like.

How Wilson, by Cartage, works with a legacy ERP's existing feed

Wilson, by Cartage, takes the order feed a legacy ERP already produces and runs the freight work from there. Wilson is an AI logistics agent. Traditional logistics software gives teams information. Wilson executes the work required to move shipments forward.

Wilson works with SAP, NetSuite, Microsoft Dynamics 365, Oracle, Infor, Epicor, Sage, Acumatica, and other ERP environments, with the exact connection depending on the customer's data flow and scope. It works from an ERP feed, a scheduled file drop, a spreadsheet, email, or an API. No integration project is required to start.

From the feed, Wilson runs quote, book, schedule, track, deliver, audit, and pay. It books only carriers the company has approved. It tracks each load using the best source available for that load. It audits each carrier invoice against what was booked and pays carriers from the Wilson Wallet. Payments above approval limits wait for a person.

Wilson replaces the TMS for freight execution. Companies without a TMS do not need to implement one. Each shipment keeps a searchable, exportable history with its change history.

Returning updates to the legacy ERP depends on scope. For how to choose which updates go back, see Freight updates back to the ERP.

Timing depends on scope: most teams start on a few lanes and expand from there. To see Wilson run freight from your current ERP export, talk to the Cartage team.

FAQs, conclusion, and sources

FAQs

Can AI automate logistics on a legacy ERP without replacing it?

Yes. Wilson, by Cartage, works from the order data a legacy ERP already sends, such as a scheduled file drop, a spreadsheet, email, or an API. The ERP stays in place as the system of record for orders.

Does a legacy ERP need an API for freight automation?

No. A scheduled file, a spreadsheet, or email can carry the orders. What matters is that the feed is complete, on time, and owned.

What happens if the scheduled order file does not arrive?

The feed owner should notice and re-run it. That check belongs in the daily routine. A missed file with no owner is the most common silent failure.

How should a corrected order be handled?

It should arrive with the same order number and update the existing shipment. If the carrier already accepted the load, the change may need a person's approval.

Will shipment updates flow back into the legacy ERP?

That depends on scope. Write-back is decided per customer. Wilson keeps the full shipment history in its own records either way.

Conclusion

A legacy ERP is rarely what blocks freight automation. The blocker is usually an unowned hand-off. Make the order feed reliable, give it an owner, and settle how corrections work. Then hand the carrier work to an AI agent. Wilson, by Cartage, runs that work from the feed the ERP already produces.

Sources