A custom ERP should not block carrier booking
Key Takeaways
A custom ERP can supply freight work through an agreed order handoff instead of blocking progress until a broad integration is built. Your AI logistics coordinator still needs complete shipment facts, readiness rules, and carrier acceptance. A daily file may be a proposed starting method. Confirm the actual scope with Cartage, including changes and result records. An available file is not proof of universal compatibility or two-way updates.
Separate order intake from booking and return updates.
Name the owner of each shared field.
Test changed, incomplete, and repeated orders.
Agree on the exact proof of carrier acceptance.
A manufacturer can keep its ERP while changing who coordinates freight. Wilson, by Cartage, is an AI logistics coordinator that replaces the TMS for freight execution. The connection method depends on the customer's environment. Buyers should define the work before choosing an interface.
The ERP integration guide explains the difference between receiving data and completing transport work. Use this article to prepare questions for operations, IT, and finance. An AI logistics coordinator works best when those teams agree on ownership before rollout.
Updated November 2, 2026
What should the connection agreement include?
The agreement should name each input, action, output, and owner. An AI logistics coordinator cannot resolve a vague ownership rule through better formatting. Put the important decisions in a short working table.
Which connection options does a custom ERP have, and what are their limits?
A custom ERP can usually connect through one of four methods. An AI logistics coordinator can start with the simplest method that delivers complete shipment facts, and the team can move to a deeper connection later if it is worth the effort.
| Option | How it works | Limits |
|---|---|---|
| Order email | The ERP or a person emails each order | Depends on consistent content; corrections need a clear rule |
| Scheduled file drop | The ERP exports a CSV or Excel file on a schedule | Changes between exports are delayed until the next file |
| Spreadsheet | Orders are kept in a shared sheet | Needs one owner and a clear change rule |
| API | The ERP sends orders through a direct connection | Needs IT time and a maintained connection |
Wilson works from the connection or data flow the company already has: an ERP feed, a scheduled file drop, a spreadsheet, email, or API. No integration project is required to start. Whether updates are written back to the custom ERP depends on the agreed scope.
| Connection area | Question to answer | Evidence to request |
|---|---|---|
| Order intake | Which orders are ready for freight? | Sample eligible order |
| Shared fields | Which system owns each value? | Field mapping |
| Booking | What proves carrier acceptance? | Accepted tender record |
| Changes | Who approves a revised order? | Change example |
| Return updates | Which events go back to the ERP? | Agreed event list |
| Failures | Who handles a failed transfer? | Named response owner |
A custom ERP may use a daily export, email, or a scoped connection. An AI logistics coordinator should not require the buyer to confuse these methods. A file received once per day does not become a continuous connection because both systems share an order number.
The order-to-freight workflow provides a useful companion. An AI logistics coordinator needs the order's transport requirements, while the ERP remains the order record. Keep that boundary visible during every technical discussion.
How can a connection agreement stay readable?
A readable agreement follows one order through the operation. Your AI logistics coordinator needs a sequence that business owners can inspect. Begin with the released order. Show the transport request, carrier response, approval where required, and acceptance evidence. Then describe any return information the ERP needs.
Separate included work from open questions in that sequence. Your AI logistics employee evaluation may confirm file intake while leaving an accounting update unresolved. Write both facts plainly. Avoid a single integration checkmark that suggests every direction and event is already supported. One completed part does not establish the rest.
Name owners by their operating role and identify the current contact separately. Your AI logistics coordinator arrangement should remain understandable when employees change. Warehouse release, IT transfer support, freight approval, and finance review may have different owners. The agreement should make those differences useful to the next person.
Describe evidence using business language. Your AI logistics employee buyer should ask for an accepted carrier commitment, a recorded revision, or a reviewed invoice. A technical success response can support the process, but it does not replace those outcomes. The agreement should explain what each piece of evidence proves.
What belongs in the unresolved-question list?
The unresolved-question list should contain decisions that still affect the proposed workflow. Your AI logistics coordinator evaluation may need answers about transfer permissions, repeat handling, or supported return events. Give each question an owner and a clear proof request. A long list without owners rarely helps the next discussion.
For example, ask how the proposed connection identifies an order revision. Your AI logistics employee needs to distinguish a changed shipment from an identical resend. The required demonstration should use both examples. Do not accept a general statement about automation as proof of either behavior.
Ask how access changes when an employee leaves or a service account changes. Your AI logistics coordinator implementation should follow the customer's security review. Specific access, logging, and credential controls need confirmation for the proposed environment. This article describes buying questions rather than asserting a particular security certification.
Finish the agreement with a boundary statement that everyone understands. Your AI logistics employee may coordinate freight while the ERP remains the business order record. The team should know which system owns each important value. Clear boundaries reduce arguments over which screen should control an operational decision.
Which order fields should IT provide first?
IT should provide the smallest complete shipment record that supports the intended workflow. An AI logistics coordinator needs enough information to ask carriers a clear question. Extra fields cannot compensate for a missing pickup address or shipment weight.
Start with the source order ID, shipping locations, requested dates, quantities, packaging, and weight. An AI logistics coordinator also needs relevant equipment, handling, appointment, and contact details. Add fields when the shipment requires them. Do not send unrelated customer or financial data by default.
A distributor might store weight at the product level but ship mixed pallets. In this illustrative example, the AI logistics coordinator needs the shipment weight. IT must explain who calculates that value. A product list without shipment totals may leave the quoting task unfinished.
Ask the team to show the same field in the export and the freight request. An AI logistics coordinator should receive the correct units and location details. A familiar column heading can still hide pounds versus kilograms or an office address versus a loading dock.
How should operations choose the first shipment example?
Operations should choose a shipment that reflects ordinary work, including ordinary uncertainty. Your AI logistics coordinator evaluation should not begin with a special demonstration order that never changes. Use an order with familiar packaging, known locations, and a normal carrier process. Remove private information where needed without removing the business problem.
Ask the warehouse to explain when that order becomes ready for freight. Your AI logistics employee needs a release rule rather than an assumption based on the order's existence. Stock may still be unavailable. Packaging may still change. The transport handoff should use the point when the company authorizes the next freight action.
Then identify the information that changes a carrier's answer. Your AI logistics coordinator may need equipment, loading method, receiving hours, and handling instructions. Ask the actual shipment owner to confirm those facts. An IT mapping can copy a field accurately while the field itself contains an outdated operating instruction.
Define the source of derived values as well. Your AI logistics employee evaluation should ask who converts product quantities into shipment weight or pallet count. That calculation may remain in the ERP, warehouse process, or another agreed step. Do not leave the responsibility hidden behind a column labeled shipment details.
Which fields should remain outside the first scope?
Fields should remain outside the first scope when they do not support the agreed freight work. Your AI logistics coordinator needs relevant order information, not every record the ERP can export. Operations and IT should explain each included field. Unused data can make the mapping harder to understand and maintain.
For example, a product margin field may have no role in an ordinary carrier request. Your AI logistics employee setup should avoid sending it merely because it is available. A receiving contact may be essential. The right distinction follows the task and customer requirements rather than the size of the export.
Keep optional fields separate from required fields. Your AI logistics coordinator evaluation should show what happens when either type is absent. A missing internal note may allow work to continue. A missing destination may block the request. The customer should approve those rules before treating every received order as usable.
Ask a second employee to read the proposed sample without help. Your AI logistics employee discussion should reveal whether the names, units, and references make sense. If that person cannot identify the shipment, the mapping needs clearer definitions. The test checks shared understanding rather than technical sophistication.
How should the team handle order changes?
The team should define which changes pause freight work and which changes only update a record. An AI logistics coordinator needs a clear rule when an order changes after quoting. A revised quantity may affect equipment, price, or the carrier's acceptance.
Choose an order version or another clear way to distinguish revisions. An AI logistics coordinator should not treat an old file as a new instruction. Ask the provider to demonstrate the proposed behavior with a repeated order and a changed pickup date.
Separate an ERP cancellation from a carrier cancellation. An AI logistics coordinator may need to contact the carrier before the freight commitment can end. Deleting a row from a file does not prove that the carrier received or accepted a cancellation.
Name the person who resolves conflicting instructions. An AI logistics coordinator should escalate when sales changes the destination but the warehouse has already released the goods. This is a business decision first. IT can support the decision only after the company defines it.
How should the team classify a changed order?
The team should classify a change by its effect on transport work. Your AI logistics coordinator needs to know whether the change affects the request, approval, carrier commitment, or internal record. A spelling correction and a new delivery address are different events. Their next actions should reflect that difference.
Use four sample changes during the design discussion. Your AI logistics employee evaluation can examine a quantity change, pickup revision, destination change, and cancellation. For each, identify the current shipment state. Ask whether the carrier has received a tender and whether it has accepted. That context determines the next question.
A change before quoting may simply require corrected shipment facts. Your AI logistics coordinator still needs the customer to confirm which facts are final. A change after acceptance may require a carrier response or another approval. The system should not silently assume the old rate and commitment remain valid.
Keep the original instruction understandable after a revision. Your AI logistics employee review should ask what the next person sees when investigating a later invoice. The buyer should confirm available revision records and document controls. Do not assume unlimited history or a particular archive behavior without a demonstrated scope.
Who decides when two instructions conflict?
A named business owner should decide when two instructions conflict. Your AI logistics coordinator cannot determine the company's priorities from competing messages alone. Sales may request a new destination while the warehouse has already loaded the original order. The freight team needs a clear decision before communicating another plan.
Write a short conflict request that identifies the order, current commitment, proposed change, and decision needed. Your AI logistics employee should help keep that question specific. Avoid forwarding an entire conversation without stating the unresolved issue. The approver needs to understand the effect of saying yes or no.
After the decision, identify each party that needs the revised instruction. Your AI logistics coordinator arrangement may include the carrier, warehouse, customer owner, and ERP owner. Confirm the communication method for the proposed setup. A changed field does not prove that every affected person has received the new plan.
Close the revision with evidence appropriate to the action. Your AI logistics employee test should distinguish a cancellation request from carrier acknowledgment. It should also distinguish a proposed pickup from an accepted appointment. Those distinctions help the company avoid treating an unresolved change as completed work.
What should return updates prove?
Return updates should describe a real freight event with a source and time. An AI logistics coordinator should not return “booked” merely because someone selected a rate. The agreed event must distinguish a request sent from a carrier acceptance received.
List the events the ERP actually needs. An AI logistics coordinator may coordinate booking, tracking, documents, and invoice review. The ERP connection may cover only some of those events. Confirm each supported update rather than assuming a complete two-way connection.
Ask how a failed return update becomes visible to the team. An AI logistics coordinator can have a valid carrier confirmation while the ERP still shows an older status. The buyer needs a recovery owner and a method to reconcile the two records.
Document links require the same care. An AI logistics coordinator may work with BOLs and PODs, but the customer should confirm where those files live. Agree on access, order references, and the handling of replacement documents before relying on ERP links.
Which freight events are worth returning?
Return the freight events that support a real business decision. Your AI logistics coordinator connection does not need a large event list merely to appear complete. The ERP owner should explain how each proposed update will be used. Booking, shipment progress, documents, and financial review may serve different teams.
For a booking event, define the evidence that supports acceptance. Your AI logistics employee evaluation should identify the carrier, shipment, and relevant commitment. The ERP should not show an accepted load solely because a request was sent. A clear event definition protects the meaning of the status.
For a delivery event, decide whether the company needs a carrier report or a verified receiving outcome. Your AI logistics coordinator may obtain available tracking information and coordinate follow-up. A reported delivery may still leave a shortage question open. The customer must define when the order can be treated as complete.
For an invoice event, distinguish received, reviewed, approved, and paid. Your AI logistics employee discussion should involve finance before selecting labels. These words describe different decisions. A document link alone does not prove approval, and a completed operational review does not prove that payment has settled.
How should teams handle conflicting status values?
Teams should identify which event and source each status represents before correcting a difference. Your AI logistics coordinator may have a recent carrier answer while the ERP displays an older update. That situation needs reconciliation. It does not automatically mean the carrier answer or the ERP value is wrong.
Use a test with a valid booking and a failed return transfer. Your AI logistics employee evaluation should preserve the real carrier commitment while the team repairs the ERP record. Ask who notices the difference and who can retry or correct the update. Those controls should be demonstrated for the actual design.
Check the meaning of timestamps and references in the returned information. Your AI logistics coordinator needs a clear connection between the event and the right movement. A timestamp may describe when a system received information rather than when the carrier performed the event. Ask which meaning applies before using it operationally.
Document links need an equally clear agreement. Your AI logistics employee setup should explain where the document lives and who can access it. Confirm the handling of revised or replaced documents. Buyers should not assume that every ERP user automatically receives access to every linked freight file.
How should buyers test a custom connection?
Buyers should test one complete order and several realistic exceptions. An AI logistics coordinator demo should show the full path from input to carrier response. A successful file upload alone proves only that the file arrived.
Use a normal shipment, an incomplete record, a repeated export, and a revised order. An AI logistics coordinator should demonstrate the agreed response to each. Record expected behavior before the test. This prevents the team from accepting an explanation instead of a result.
Include a handoff test for the warehouse. An AI logistics coordinator should help the team know which carrier accepted and what remains unconfirmed. The warehouse needs reliable instructions, not a technical success message from the integration tool.
Finish with a simple reconciliation exercise. An AI logistics coordinator and the ERP team should identify the same order, shipment, carrier, and status. Any difference should have an owner. The purpose is dependable freight work, not a large list of connected fields.
How can a buyer create a useful test pack?
A useful test pack contains normal work and exceptions that change the next action. Your AI logistics coordinator evaluation should use the customer's actual shipment pattern. Include a clean released order, an incomplete record, a repeated export, and a revision after acceptance. Keep expected results beside each example.
For the clean order, require a complete path to carrier acceptance. Your AI logistics employee demonstration should identify what the connection receives and what freight coordination follows. Record any manual steps honestly. A person approving a quote may be an intentional control rather than an unfinished technical feature.
For the incomplete record, remove a fact that matters to the shipment. Your AI logistics coordinator test should show who requests it and where the order waits. Do not award a pass merely because the software accepts the row. A usable freight request needs enough information for the carrier's actual decision.
For the repeated export, ask the provider to explain the proposed repeat-handling behavior. Your AI logistics employee evaluation should then show that behavior with the same order. Include a true revision afterward. This checks whether the design distinguishes repeated information from changed instructions.
What should the review team record during the test?
The review team should record observed actions, missing evidence, and remaining ownership questions. Your AI logistics coordinator score should follow what the provider demonstrates. Use plain outcomes such as complete, incomplete, or needs clarification. A weighted score can hide an essential missing step if the team focuses only on totals.
Give warehouse, finance, operations, and IT different questions to inspect. Your AI logistics employee evaluation should ask the warehouse whether the pickup instructions are usable. Finance can inspect references and the later review handoff. IT can inspect transfer scope and recovery. Operations can inspect carrier communication and acceptance.
Test recovery before discussing expansion. Your AI logistics coordinator workflow should explain how work resumes after a rejected record or failed return update. The team needs to know the shipment's current state before repeating an action. A recovery explanation should identify who acts and what evidence they check.
Finish with an open-issue review rather than a forced success statement. Your AI logistics employee buyer should distinguish demonstrated work from proposed follow-up. Assign unresolved items to the right owner. The purpose is a dependable operating agreement, not a demonstration that appears perfect because exceptions were removed.
What should the team approve before expanding?
The team should approve the field mapping, decision rules, and recovery process before expanding the connection. An AI logistics coordinator needs a stable operating agreement. A successful demonstration does not settle every future order type.
Name the person who approves changes to the mapping. An AI logistics coordinator may receive new fields or revised business rules later. IT and operations should review those changes together so a technical update does not create an unintended freight instruction.
What should remain stable before more order types are added?
The team should stabilize the meaning of the first order flow before adding more complexity. Your AI logistics coordinator needs agreed fields, release rules, change handling, and carrier acceptance evidence. If those basics remain unclear, adding more order types can multiply confusion. Expansion should follow understanding rather than a larger import count.
Review the first examples with the people who perform the actual work. Your AI logistics employee setup may look clear to a project owner but remain confusing to the warehouse. Ask each role to explain its next action. If the explanation differs between people, revisit the agreement before adding another workflow.
Keep changes to mapping under a named review process. Your AI logistics coordinator connection may receive revised location codes, packaging values, or order statuses later. IT and operations should check the effect together. A technical adjustment can change which shipments become eligible for carrier coordination.
Define the boundary between operational feedback and product scope. Your AI logistics employee evaluation may reveal a useful requested improvement. Record it separately from work already agreed. Do not describe a possible enhancement as a capability the current setup includes. The commercial and technical scope should stay consistent.
How should leaders judge the custom connection?
Leaders should judge the connection by whether it supports understandable freight work with clear ownership. Your AI logistics coordinator should help move a released order toward a confirmed carrier commitment. The connection is useful when the relevant teams can follow that process and resolve exceptions. A large field count is not enough.
Review the work that still requires a person and why it remains. Your AI logistics employee may need approval for an expensive load or conflicting instruction. Those decisions can be appropriate. Distinguish necessary authority from avoidable copying, chasing, and rebuilding information. That distinction makes the buying decision more practical.
Use the company's own observations to discuss improvement. Your AI logistics coordinator review can record incomplete orders, unresolved requests, accepted bookings, and recovery work. These categories do not establish a promised result or industry benchmark. They give leaders a clear way to examine the proposed workflow.
The final decision should name the supported order flow and its limits. Your AI logistics employee arrangement should make the first operating step clear. Confirm the input, carrier work, approval owner, and agreed return information. That is a stronger buying foundation than a promise that a custom ERP simply connects.
What should a file-based evaluation prove first?
It should prove that a representative order becomes an accurate shipment and a clear carrier commitment. Your AI logistics coordinator test should include a missing field and a changed record. Inspect the source relationship, approval owner, and final booking position. This is a buyer-controlled evaluation, not a promise of a standard commercial pilot, free trial, or fixed implementation schedule. Agree any commercial evaluation terms separately.
Your AI logistics employee demonstration should also show which result the ERP users can find afterward. A message, file, shared record, or API update can represent different scopes. Name the actual method and any remaining owner. A clean import does not establish a tested return path. The business should decide whether the proposed starting workflow meets its immediate need without pretending that every later integration requirement has already been completed.
FAQs
How can a custom ERP support carrier booking without a broad API project?
Define an agreed input such as a supported file, the required shipment facts, and the permitted carrier workflow. Your AI logistics coordinator should demonstrate the resulting booking and record handoff. Confirm the actual method for the customer's environment. Keep later return updates separate from initial order intake.
How can a custom ERP connect to freight software, and what are the limits?
Through order emails, a scheduled file, a spreadsheet, or an API. An AI logistics coordinator can start with the simplest method that gives complete shipment facts; file and email methods are slower to reflect changes, and write-back depends on the agreed scope.
Does Wilson require us to replace our ERP?
No. Wilson is an AI logistics coordinator for freight execution and connects to the customer's order workflow. The exact intake and return methods depend on the environment. Confirm those methods for your ERP.
Can we start with a file instead of an API?
A file-based start can be appropriate when its contents and timing support the workflow. An AI logistics coordinator still needs change rules and complete shipment information. Confirm the specific file process with Cartage.
Does receiving an ERP order mean the load is booked?
No. An AI logistics coordinator must still complete the relevant carrier coordination. The booking record should show acceptance rather than only order receipt or rate selection.