# Building a custom TMS in-house: what to plan after launch

[Source](https://onechain-tms.com/en/blog/developper-son-tms-en-interne/)

Language: en

Author: OneChain

Published: 2026-11-18

Updated: 2026-11-19

**Targeted development can make sense when the need is clear, the scope is narrow and the technical team has time for it. After that, you own the partner connections, the business rules, user support and every future change. Building one function, such as printing labels through an express carrier’s API, is a very different commitment from building a complete TMS.**

The question of building a TMS in-house comes up more often now that AI tools can speed up technical teams, and it is a fair question to ask. This article covers what to examine before deciding, especially what weighs on you once the tool is live.

## When does building in-house make sense?

When three conditions are met. The business need is clearly stated and has been discussed between transport and technical teams. Developers are available for the project, not just to launch it. The scope is narrow.

If one is missing, the risk changes. A vague need ends up in the code just as vague. A team busy elsewhere delivers a first version and then cannot develop it further. A broad scope turns an application into a full system.

AI tools can help write code faster. They do not shorten the list of what you will have to own once the tool is running, and that list is what the decision should rest on.

## A targeted function or a complete TMS?

They are two different projects. Connecting your system to an express carrier’s API to print labels is a targeted need: one partner, one flow, and a result that is easy to check.

A complete TMS covers far more. It handles the orders to be shipped, rate cards, pre-invoice calculation, allocating shipments to providers and filling equipment such as containers or trucks. Each of these has to be designed, built, tested and maintained. Before estimating anything, write down the functions the tool will need to cover over time, not just at launch.

## Who will connect your forwarders and carriers?

Your team, partner by partner. Forwarders and carriers supply part of the data: bookings, dates, documents, invoices. For them, a tool built by one shipper is one more system to connect to or learn.

A solution already used by other shippers may be familiar to some partners’ teams, which makes exchanges easier without guaranteeing an immediate connection. With an in-house tool, every new partner, every format change and every failed transmission lands with your team.

Not every partner has an API. Plan for other routes: direct access to the tool, document exchange, and integration where it exists. Our [freight forwarder network](https://onechain-tms.com/en/freight-forwarder-network/) page describes these three ways of working with a partner, and the [Integrations](https://onechain-tms.com/en/integrations/) page lists what to define for each exchange: fields, direction, channel and who owns errors.

## Who will maintain the business rules?

A pair, one on the business side and one on the technical side. Rules change: a new rate card, a new surcharge, a new allocation rule, a new site. Each time, the rule has to be translated, tested and checked against the ones already in place.

For each rule, name a business owner who approves changes and a technical owner who makes them. Without those two names, the rule ends up living in a spreadsheet next to the tool.

## Who will handle support and future changes?

An internal product team: once in service, the tool becomes a product to maintain. Once the tool is live, users report issues, ask for changes and expect answers. Partner APIs change, technical dependencies need updating, and the people who built the tool sometimes move on.

Before development starts, answer a few questions. Who receives user requests? Who decides between two competing changes? Who takes over the code if its author leaves? If those answers are unclear, the real cost of the project is not yet known.

## How can in-house development and vendor software work together?

By deciding what you want to build and maintain yourself, and what you would rather hand to a vendor. Thinking through an in-house build is useful in itself: it gets technical and business teams talking, clarifies processes and helps you choose a solution that fits the way you work.

| Responsibility | If you keep it in-house | If a vendor carries it | Question to ask |
|---|---|---|---|
| Partner connections | Your team negotiates and maintains each exchange | Ask who defines and maintains each exchange | How many partners and formats will we have later? |
| Freight business rules | You code and test them | Check which ones can be configured without development | Are our rules really different from common practice? |
| User support | Your team answers | Have support terms written into the contract | Who answers when a shipment is blocked? |
| Future changes | Your own roadmap | Ask how your requests are taken into account | Which changes genuinely set us apart? |
| Company-specific functions | Targeted in-house development | Check which data the vendor exposes to your tools | Which data do we need to read or write? |

The split does not have to be all or nothing. A company can rely on a vendor for transport management and keep dashboards, checks or agents that read its data in-house. In that case, check what the vendor exposes. At OneChain, API integrations and headless approaches (using OneChain’s data and operations from your own tools, without going through its interface) are designed to connect data to internal tools, such as a reporting (BI) tool or an AI agent; the shipments, fields and operations made available are defined for each project. The [Tools and agents](https://onechain-tms.com/en/headless/) page describes these access methods.

If the hesitation is really about your ERP’s transport module, our article on [ERP transport modules versus specialist TMS](https://onechain-tms.com/en/blog/module-transport-erp-ou-tms/) offers a comparison grid. To place this decision among the others, see [the questions to settle before starting a TMS project](https://onechain-tms.com/en/blog/cadrer-un-projet-tms/).

To discuss your own case, [set up a conversation with the team](https://onechain-tms.com/en/contact/?sujet=Build%20or%20buy%20a%20TMS) and describe the need, the functions you are considering building and your partners. Keep personal data and confidential documents out of the form.

