Which ERPs does Wilson integrate with (NetSuite, SAP, Oracle, Epicor, proprietary systems)?
Key Takeaways
Updated October 28, 2026
Wilson can work from ERP order data through a customer-specific setup. The method can use files, email, spreadsheets, or an API connection. Your AI logistics coordinator does not require the same integration method for every ERP. Confirm the proposed data flow for your NetSuite, SAP, Oracle, Epicor, or proprietary environment before treating it as a ready-made connector.
- Start with the freight job and the order data it needs.
- Separate order intake from updates sent back to the ERP.
- Confirm the system version, data method, and responsible team.
- Test a changed order as well as a new order.
- Follow the test through the relevant carrier action.
- Keep planning, inventory, and accounting responsibilities clear.
Wilson, by Cartage, is an AI logistics employee for freight execution. Wilson turns usable order information into coordination work across quoting, booking, tracking, documents, audit, and payment. The ERP can remain the source of the business order. The integration connects that record to the freight work without implying that Wilson replaces every ERP function.
What does an ERP integration with Wilson actually mean?
An ERP integration supplies the order information needed for an agreed freight workflow. Your AI logistics coordinator may receive that information through an export, email, spreadsheet, or API. The setup should define which data moves, when it moves, and what action it permits. A software brand name alone does not define those details.
| ERP environment | Useful starting question | What needs customer-specific confirmation |
|---|---|---|
| NetSuite | Where does the released shipment instruction come from? | Account setup, fields, input method, and any return updates |
| SAP | Which order or delivery record starts the freight task? | System environment, release rules, interfaces, and ownership |
| Oracle | Which application and record hold the relevant order? | Product, version, data path, and the intended freight scope |
| Epicor | Which export or interface supplies the transport details? | Available information, layout, change handling, and setup work |
| Proprietary system | Can the system provide a reliable order input? | File, email, spreadsheet, or API arrangement and technical owner |
| Mixed environment | Which source governs each shipment instruction? | Shared references, handoffs, exceptions, and reconciliation |
Distinguish compatibility from a packaged connector
A customer-specific data arrangement differs from a connector that is ready for every account. Your AI logistics employee can be evaluated through the actual order path available to your business. The buyer should ask what already exists and what needs configuration. This article does not claim a verified production connector directory for each named ERP. It explains the workflow questions that establish a useful connection.
Consider an illustrative distributor of industrial valves. Its ERP stores customer orders, while employees request carrier quotes by email. An AI logistics coordinator needs the transport instruction for a released shipment. The first question is which record supplies that instruction. The next question is how changes reach the workflow. Naming the ERP does not answer either question by itself.
Name the application and environment
The implementation discussion should identify the actual system, rather than a broad product family. Your AI logistics employee setup may depend on the application, version, permissions, and existing interfaces. Two businesses using the same ERP brand can have different order layouts. The provider should review the environment that will supply the data. A general list of familiar logos is not enough.
The valve distributor should bring a representative order and explain where its freight details live. Its AI logistics coordinator may need facts from more than one screen or export. The business should identify those sources before promising a simple handoff. Some fields may belong to the warehouse or customer service. That operating information matters alongside the technical connection.
Define the beginning and endpoint
The integration scope should say what starts the workflow and what counts as completion. Your AI logistics employee might begin with an order released for transport. The required endpoint may be carrier acceptance, delivery evidence, or invoice reconciliation. Those scopes need different information. A successful order transfer proves the input arrived. It does not prove that the full shipment process finished.
The distributor can define its first scope as a released order through an accepted carrier booking. Its AI logistics coordinator then needs enough information to request the right service. The team should identify what remains outside that first scope. Optional return updates can be considered separately. This makes the proposed setup understandable to operations and IT before either team estimates its own work.
What order data does Wilson need from the ERP?
Wilson needs the shipment facts required to perform the agreed task. Your AI logistics coordinator generally needs a clear reference, locations, readiness information, freight details, and service instructions. The exact fields depend on the mode and workflow. The team should map each required fact to its source rather than assume every order contains a complete transport request.
Identify the movement separately from the sales order
A sales order can create more than one shipment. Your AI logistics employee needs a reference that identifies the movement being quoted or booked. A partial release may have a different ready date from the remaining goods. The team should explain that relationship during setup. Otherwise, the order number can be too broad to identify which freight instruction the carrier has accepted.
For the valve distributor, one customer order may leave from two warehouses. Its AI logistics coordinator needs the relevant origin, quantities, and required service for each movement. The business should also explain how those shipments remain connected to the customer order. The goal is a coherent operating record. This example does not claim a particular automatic split-order feature or universal ERP mapping.
Supply the physical freight details
The transport request should describe the freight well enough for the selected service. Your AI logistics employee may need pallet count, weight, dimensions, packaging, or equipment requirements. The necessary facts vary by mode and shipment. The company should identify who owns their accuracy. A technically valid export can still be unsuitable for quoting if it omits a material shipment requirement.
The distributor should distinguish a product's stored information from the packed shipment details. Its AI logistics coordinator may receive a SKU description while the warehouse determines the actual pallet count. The setup should explain when that fact becomes available. It should also define what happens if the order reaches the workflow first. Missing packing information should remain an open question rather than an invented value.
Include dates with a clear meaning
A date field should say whether it means order creation, goods readiness, requested delivery, or another event. Your AI logistics employee needs that meaning to coordinate the correct service. A bare date can create confusion when different teams use it differently. The mapping discussion should resolve the ambiguity. The same principle applies to time windows, site hours, and relevant operating calendars.
The valve distributor may record a requested ship date before the goods are ready. Its AI logistics coordinator should receive a transport instruction that reflects the approved release. The business should explain whether another person confirms readiness. The integration should support that agreed process. It should not turn a planning date into a pickup instruction simply because the field is available in an export.
Keep customer instructions tied to the shipment
Some transport requirements come from the customer's order or receiving site. Your AI logistics employee needs those instructions when they affect carrier selection, booking, or communication. Examples can include a delivery appointment requirement or a named receiving contact. The company should identify where the current instruction lives. It should also explain who can approve a change when the required service cannot be met.
The distributor should test an order with a special receiving requirement. Its AI logistics coordinator needs to carry that requirement into the relevant carrier conversation. The team should inspect the resulting request and booking evidence. A copied customer note is useful only if the freight workflow applies its meaning. The buyer should evaluate the actual result rather than assume that transferring the field completes the requirement.
Can Wilson start without a full API integration?
A daily order file and a full connection are different setups, and many companies start with the first.
| Daily order file | Full connection | |
|---|---|---|
| How orders arrive | A scheduled export by file drop or email, or a spreadsheet | An API or ERP feed that sends orders as they are released |
| Speed | Orders arrive at set times, such as once or several times a day | Orders arrive as soon as they are released |
| Setup | Uses an export the company can already produce | Needs access, field mapping, and IT time |
| Updates back to the ERP | Usually a return file, if needed | Can return booking and status details, depending on scope |
| Good fit | Starting on a few lanes, or ERPs without an easy API | High volume, or teams that need status in the ERP quickly |
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. Write-back to the ERP depends on scope. No integration project is required to start.
Yes. Wilson can start from an agreed email or file-based order workflow when it supplies the needed information. Your AI logistics coordinator does not need a full API project for every starting scope. The company still needs a reliable input and clear rules for changes. Files reduce neither the need for accurate instructions nor the need to define ownership.
Choose a method that the business can maintain
The best starting input is one that the team can supply reliably for the agreed task. Your AI logistics employee may work from an existing export while a deeper connection is considered. The buyer should record any recurring preparation work. A useful first method can still require a person to release or prepare information. That responsibility should be visible in the operating plan.
Cartage's published Musser Biomass case describes an email and CSV workflow. An AI logistics coordinator can therefore be evaluated through that kind of starting point. The case does not establish that every file layout works unchanged. It also does not transfer that customer's outcome or implementation experience to another company. Bring your own example and confirm the proposed method.
Separate the input method from the quality of the instruction
An API can carry incomplete information, and a file can carry a complete transport request. Your AI logistics employee needs the right facts regardless of the connection type. The team should first check what the input means. It should then choose the appropriate technical method. Neither a modern interface nor a familiar spreadsheet resolves an unclear release instruction on its own.
For the valve distributor, a daily export may be useful for some routine movements. Its AI logistics coordinator may need a different arrangement for urgent changes. The company should explain how quickly those changes matter to operations. The setup discussion can then address the relevant data path. This article does not prescribe one transfer schedule or claim that every file-based process suits every shipment.
Agree on repeated and corrected inputs
The setup should establish what a repeated file or corrected row means. Your AI logistics employee needs to distinguish a new instruction from an update to an existing movement. The buyer should ask the provider to demonstrate that distinction in the proposed workflow. Do not infer a particular duplicate-prevention mechanism from the mere ability to read CSV or Excel files.
The distributor can submit an authorized sample twice during evaluation and ask what happens. Its AI logistics coordinator test should inspect the resulting shipment work. Then the team can change one relevant field and inspect the response. This is a buyer test, rather than a claim about a standard technical control. The expected result should be agreed before the process handles live transport instructions.
Decide when a deeper connection becomes useful
A deeper integration can be considered when the current handoff leaves important recurring work. Your AI logistics employee arrangement should be evaluated through that remaining workload. The team might need more timely input, a better change path, or return information for another system. Describe the actual problem before specifying the technology. The next integration step should serve an operating need that the business can explain.
For a separate review of the starting options, read how Wilson can work from email and order files. An AI logistics coordinator setup should fit the agreed freight task. The company should compare the proposed input, manual preparation, and exception path. That makes the decision more useful than treating an API connection as an automatic measure of success.
Does Wilson send updates back to NetSuite, SAP, Oracle, or another ERP?
Return updates depend on the customer-specific setup and agreed scope. Your AI logistics coordinator should not be presented as writing every shipment event into every ERP by default. The company should define the required return information and the supported method. Order intake and ERP write-back are separate implementation decisions, even when both belong in the eventual workflow.
Can booking details be returned to the ERP by API?
They can be when the customer's scope includes that write-back and the ERP accepts the update. An AI logistics coordinator should not be presented as writing every booking field into every ERP by default. The fields, the method, and the direction are agreed during setup and tested before go-live.
| Booking detail | Why the ERP might need it | Confirm during setup |
|---|---|---|
| Carrier name | Shipping and customer service reference | Which field receives it |
| Booked rate | Freight cost on the order or shipment | Whether finance wants the booked or the audited amount |
| Pickup date and window | Planning and customer promises | Which date field it updates |
| PRO or tracking number | Customer updates and tracking | Where it is stored |
| Status changes | Order status for sales and service | Which events count as status changes |
A PRO number is a tracking number a carrier gives to a shipment.
What should happen if a return update fails?
The setup should say what happens when an update does not reach the ERP: who is told, how the update is retried, and where the team can see what was sent. A failed update should never leave the ERP silently out of date. Testing one failed update before go-live is part of a sound ERP test.
Choose which return information has a business use
The company should identify who needs the update and what they will do with it. Your AI logistics employee might produce information relevant to customer service, receiving, or finance. Those users may need different details at different stages. The return flow should serve a clear purpose. Sending every available event back to a system can add complexity without answering the user's actual question.
The valve distributor's customer service team may need a confirmed booking reference. Its AI logistics coordinator may also obtain later delivery evidence. The company should decide where each belongs and which status it supports. A carrier's acceptance and a completed delivery are different events. The return mapping should preserve that difference rather than compress every stage into a general “shipped” label.
Confirm the method and direction
The setup should state how the return information reaches its destination. Your AI logistics employee arrangement may involve an agreed file, a shared data destination, or an API where available. The actual method must be reviewed for the environment. The provider should explain what is included. This guide does not verify a standard bidirectional connector for every named system or edition.
IT should identify the system that receives the update and the permissions required for that step. The AI logistics coordinator workflow also needs operations to explain the event's meaning. A technical field mapping can be correct while the business interpretation is wrong. Review both parts together. The test should show what the next user sees and whether that information supports the intended decision.
Distinguish a carrier event from an order decision
A freight event should not automatically be treated as a complete business-order outcome. Your AI logistics employee may receive a delivered status while the receiving team still checks a shortage. The company should define how that information affects its order record. Some decisions remain with people or another system. The integration should preserve the business's responsibility for the relevant order status.
The distributor can test a delivery event with an unresolved document or quantity question. Its AI logistics coordinator should provide the available evidence within the agreed scope. The business should then explain what its ERP status means. This evaluation does not assume a standard shortage-control feature. It checks whether the proposed handoff supports the operation's own definition of completion.
Keep accounting responsibilities clear
A freight payment and an ERP accounting entry can involve different records and responsibilities. Your AI logistics employee handles freight audit and payment coordination within the agreed arrangement. The company should confirm how the relevant finance information reaches its accounting process. Do not assume that a carrier payment automatically completes every ledger, tax, approval, or reconciliation task in the ERP.
Finance should identify the evidence it needs for its own review. The AI logistics coordinator's record may need to be connected to the shipment and supplier references that finance uses. Ask which return information is included and who handles remaining steps. This keeps the integration discussion focused on the actual work. It also avoids presenting a freight workflow as a replacement for the whole finance system.
How should IT and operations test the proposed ERP workflow?
IT and operations should follow one representative order through the agreed freight endpoint and one important exception. Your AI logistics coordinator test should inspect the input, permitted action, carrier response, and return information. A connection test alone is too narrow. The business needs to see whether the full handoff supplies useful information and completes the intended task.
Prepare a representative order
Use an authorized example that reflects the normal shipping process. Your AI logistics employee evaluation should include the real kinds of fields and service requirements the workflow will receive. Avoid a simplified record that removes the operation's main difficulty. The buyer should know what the proposed setup needs before the test. Record any preparation that a person must perform.
The valve distributor can choose an order with a normal warehouse release and an approved carrier list. Its AI logistics coordinator demonstration should show how that input becomes a useful carrier request. The team should inspect the movement reference and service details. It should also check what the carrier has actually accepted. A selected rate on a screen does not establish an accepted booking.
Add a meaningful order change
Change an instruction that matters after the first handoff. Your AI logistics employee test might use a revised ready date or a changed quantity. The team should inspect how the proposed workflow receives that change. It should then establish whether the carrier needs a new answer or approval. The test should expose the communication and decision work, rather than stop at an updated data field.
A changed record can be easy for IT to observe but difficult for operations to resolve. The AI logistics coordinator may need to contact a carrier about the revised service. The company should ask what continues and what waits for a person. The buyer should not assume a universal automatic change or cancellation process. Define the expected behavior for the actual scenario under review.
Test missing information and unclear authority
The evaluation should include an input that cannot safely support the next task. Your AI logistics employee may need a missing weight, contact, or service instruction. The setup should identify the question and responsible person. Inspect what that person receives. An unclear escalation can leave the same manual investigation in place, even when the order arrived successfully.
Then test a decision outside the agreed approval boundary. Your AI logistics coordinator should follow the customer-defined process for that situation. The provider should explain which parts are configured and what the customer still needs to do. Cartage doesn’t run pilots; most customers start Wilson on a few lanes and expand from there. It is a way to make the proposed operating scope concrete.
Review the result with the next user
The user who depends on the handoff should inspect the result. Your AI logistics employee may supply information that another team needs to act on. Customer service should understand the shipment update. Finance should understand the relevant evidence. IT should understand where the data came from and where it went. A useful review includes these perspectives without confusing their different responsibilities.
The distributor should record unresolved requirements rather than hide them in a successful connection report. Its AI logistics coordinator evaluation can then show the work completed and the work remaining. That distinction helps the buyer decide whether the proposed scope addresses the original freight problem. It also gives the implementation discussion a clear list of facts that still need an owner.
What should we bring to the first ERP integration discussion?
Bring a sample order, an explanation of its fields, the current freight process, and the desired endpoint. Your AI logistics coordinator setup also needs the people who own system access and operating decisions. The discussion should identify a workable starting scope. It should not promise a universal timeline, fixed fee, or connector before the environment is reviewed.
Prepare a simple input and output description
Write what starts the work and what the next person needs afterward. Your AI logistics employee arrangement should make that handoff understandable without specialized integration language. Identify the order source, movement reference, required action, and relevant return information. Add the person responsible for each unresolved fact. That short description can guide a more detailed technical discussion.
The valve distributor might describe its goal as turning a released warehouse order into an accepted carrier booking. Its AI logistics coordinator needs the approved carrier rules and the facts required to request that service. The company can separately list return information it wants in the ERP. This keeps the first workflow clear while leaving later requirements visible for scope review.
Bring the people who know the exceptions
The experienced coordinator often knows which apparently simple orders need another question. Your AI logistics employee setup benefits from that operating knowledge. Ask the team to bring a normal example and a difficult example. The difficult case might involve a split release or an incomplete site instruction. The purpose is to reveal the actual decision, rather than document every rare event before starting.
IT should also explain existing exports, interfaces, and access restrictions. The AI logistics coordinator discussion needs those technical facts alongside the operating playbook. If a required method is unavailable, the team can review another proposed path. That is different from promising compatibility without inspection. The business should leave with a named next step for both the data handoff and the freight task.
Keep the replacement scope separate from the ERP connection
Wilson replaces the TMS for freight execution; it does not replace the entire ERP. Your AI logistics employee can work from order information while the ERP retains its wider business role. Planning, inventory, and other responsibilities still need clear owners. Read how Wilson differs from a TMS when defining that boundary.
The final proposal should explain the input, freight work, approval rules, and return path in plain language. Your AI logistics coordinator arrangement should also identify customer-specific setup and remaining manual work. Ask for clarification where a broad integration claim hides those details. A clear operating scope is more useful than assuming the same connector, permissions, or data flow applies to every company.
Ask for a written description of the proposed data path
A short written scope should name the input, timing, owner, and intended freight action. Your AI logistics employee setup should also identify any return information that belongs in the first stage. The description does not need complex language. It needs to explain what the company will provide and what the proposed workflow will do with it. Ask the provider to identify assumptions that still need checking.
The valve distributor can review that description with operations and IT together. Its AI logistics coordinator may depend on a warehouse release that IT does not see in the normal export. The discussion should resolve who supplies that fact. If a return update remains outside the initial scope, say so clearly. This helps the business plan the remaining handoff without treating an unconfigured feature as part of the implementation.
The same record can guide the later acceptance review. Your AI logistics employee test should follow the agreed input and inspect the promised endpoint. Record any difference between the proposal and the observed result. The point is to understand the actual operating arrangement before relying on it for live shipment work.
For a step-by-step example with Acumatica, see how an Acumatica sales order becomes a load accepted by an approved carrier.
FAQs
Which ERPs does Wilson integrate with (NetSuite, SAP, Oracle, Epicor, proprietary systems)?
Wilson can work from ERP order data through a customer-specific method. Your AI logistics coordinator setup should confirm the actual NetSuite, SAP, Oracle, Epicor, or proprietary environment. Do not assume every system has the same ready-made connector.
Can Wilson return booking details to the ERP by API?
It depends on the customer's scope and the ERP. An AI logistics coordinator can return agreed booking fields, such as carrier, rate, and pickup date, when the setup includes that write-back; the fields, method, and failure handling are confirmed during setup.
Do we need an API before we can start?
Not always. Your AI logistics employee can work from an agreed email, spreadsheet, or file-based input when it supplies the information needed for the selected workflow.
Does reading ERP orders include automatic write-back?
No automatic assumption is appropriate. Your AI logistics coordinator setup should define any return information and the supported method separately from order intake.
What should we test before relying on the connection?
Test a representative order, a meaningful change, and missing information. Your AI logistics employee evaluation should follow the relevant carrier action and any agreed return update.