TMS automation that fills the system you already pay for

TMS automation is software that reads the documents arriving in a freight inbox and writes their fields into the transport management system you already run, so a person stops retyping them. It adds no new system to log into. The TMS stays where it is; only the typing moves.

Key takeaways
  • TMS automation fills the system you already pay for from the documents that arrive. It does not replace the system, and a project that begins by replacing one rarely ends.
  • A TMS sits half empty because it stores what a person types, and nobody is paid to type. The gap is a labour problem, not a software defect.
  • An agent writes by whatever the system accepts, which is an import file, an API, or the same screens a person would use. Which of these is available depends on each installation.
  • Under 49 CFR 371.3 a broker must keep a record of each transaction for three years, so the fields an agent writes are also the record the regulation expects to exist.
  • Before signing with any vendor, ask what it writes, by which route, what it does with a document it cannot read, and what you can take with you if you leave.
The answer

What is TMS automation?

TMS automation is software that reads the documents arriving in a freight inbox and writes their fields into the transport management system a company already runs. It does the keying that a dispatcher, an account manager or an accounts-receivable clerk would otherwise do between calls. The documents are the familiar ones: rate confirmations, bills of lading, proofs of delivery, carrier invoices. The system is whichever one you already pay for.

The phrase gets used loosely for anything that touches a TMS, including route optimization and load matching. This page means the narrow thing: moving information from documents into the system without a person retyping it. The broader picture is in logistics automation.

Why is a TMS still empty when you are paying for it?

Because a TMS records what someone types, and nobody is paid to type. Typing is the work that loses to the next phone call. A load number gets entered because the load cannot move without it. A time at the dock, a note about a damaged pallet, or the scanned proof of delivery does not stop anything that day, so it stays in an email thread, a driver text or a folder on someone’s laptop, and the system holds a skeleton of the shipment instead of the shipment.

This is why buying a better TMS does not close the gap. The new system will be as empty as the old one for the same reason, plus a few months in which both are partly filled. The gap sits in the labour, and the labour is the thing the software is waiting on. Where the missing fields are facts that no document ever held, reading will not help either; that is covered under freight document automation.

What gets written into the TMS, and by which route?

The agent writes the fields of each document into the record they belong to, by whichever route your system accepts: an import, an API, or the same screens a person would fill. We do not claim an integration with any named product. What a system exposes depends on the plan, the version, and what the vendor has chosen to switch on, and two customers of one product can differ. So the route is found by looking at your installation, and the table sets out what each one means in practice.

RouteWhat it isWhere it fits
ImportA file in a format the TMS reads, such as a spreadsheet or delimited text, loaded on a scheduleBatches of loads or invoices where a delay of a few hours is acceptable
APIA direct connection where the agent sends each record as it is readyOperations that need the record current within minutes, if the vendor offers access
The screensThe agent fills the same forms and fields a person would, by driving the interfaceSystems with no import or API, accepting that it is the slowest and the most fragile

Whatever the route, the first thing to settle is what happens to the original page. A written field with no link to its source is a number that someone will eventually doubt and cannot check.

What does the record have to hold?

At minimum, the transaction details that a broker is required to keep for three years, which is the line between a database and a file. 49 CFR 371.3 requires a broker to keep a record of each transaction, and it names what goes in it: the consignor’s name and address, the originating carrier’s name, address and registration number, the bill of lading or freight bill number, the compensation the broker received and who paid it, any non-brokerage service performed, and the freight charges the broker collected with the date the carrier was paid. A TMS that was filled by hand holds this only for the loads where someone found the time.

The rule is a floor and not a design brief, but it changes how the question reads. A broker cannot treat a half-empty system as a convenience that might be tidied up later, because the entries are the record the regulation expects to exist. Automation that writes every load the same way produces that record as a by-product. On the carrier side, the invoice that follows has its own content requirements under 49 CFR 373.103, and a field written from the document that proves it is the one that survives a question.

What should you ask a TMS automation vendor before signing?

Ask what it writes, by which route, what it does with a document it cannot read, and what leaves with you if you cancel. The first two questions are easy to answer and tell you little, since every vendor will say it writes to your system. The third is the real one. A tool that stops at the first unreadable scan has moved the work to a queue that nobody owns, and a tool that guesses has hidden it. Ask to see a held item and the person it was routed to.

The fourth question is about leaving. If the values the agent wrote live in your system, they stay when it goes. If they live in the vendor’s own platform, with your TMS holding a copy, you have rented your own records. Also ask whether the page stays linked to each field, whether pricing is by document or by seat, and who answers when the writing route breaks after a TMS update, because it will.

Do you have to migrate to a new TMS?

No, and a project that starts by replacing the TMS usually does not finish. A migration means moving years of loads, customers, rates and carrier history, retraining every user, and running two systems while the old one winds down. It competes for the same few people who do the daily work, and it is the most common reason that back-office improvement is abandoned halfway. Automation does not need it and does not delay it.

If you do have a reason to move later, the work carries over. The documents are still the documents, the capture process still holds, and the agent points at a different place to write. If the current system is not the problem, the problem is the typing, and that is the part you can remove this quarter without touching the contract.

By Deivy Stiven Hernandez, founder at catuh. Published . Last revised . Filed under Automation.

Step by step

How to automate a TMS without replacing it

Five steps, none of which touch your contract with the TMS vendor. The first two are an audit you can do alone in a week, and they tell you whether the rest is worth building.

  1. 01 / 05

    List the fields your team retypes from documents

    Walk one load from rate confirmation to paid invoice and write down every field that was typed from a page. That list is the scope. It is usually longer than the screens suggest, because some of it lives in spreadsheets beside the system.

  2. 02 / 05

    Find out how your TMS accepts data

    Ask the vendor, or read the documentation, for the three routes in order of reliability. An import file, an API, and the user interface. Write down which exist for your plan, since that varies by installation and by what you have licensed.

  3. 03 / 05

    Mark the fields that are blank rather than mistyped

    Arrival and departure times, the exception at pickup, the carrier check. No agent can read these from a page, so decide who records each one and when, before building anything else.

  4. 04 / 05

    Run the agent beside the person for a short period

    Both write the same load. Compare the two records field by field and look at what the agent held back. The comparison shows where it is reliable and which document types need a person.

  5. 05 / 05

    Keep the page linked to the field

    Every written value should point back to the page it came from, so a disputed field can be traced in a minute. If that link cannot be kept inside your TMS, keep it beside it.

Questions

Frequently asked questions

What is TMS automation?

TMS automation is software that reads the documents a freight company receives and writes their fields into the transport management system the company already uses, so nobody retypes them. It covers rate confirmations, bills of lading, proofs of delivery and carrier invoices. The system itself is unchanged; what changes is who enters the data, which becomes a program that holds back what it cannot read.

Do I have to change my TMS to automate it?

No. Replacing a TMS is a long project with its own risks, and automation does not depend on it. The agent writes into the system you have, through an import, an API or the same screens a person would fill. A migration can still be worth doing for other reasons, but it should not be the price of getting the paperwork off your team.

Why is my TMS empty even though we pay for it?

Because a TMS records what is entered, and entering is unpaid work that loses to the next call. Load numbers get typed, while arrival times, exceptions and the documents themselves are left in email. The software is not failing. It is faithfully holding the small share of the paperwork that someone had time to key in.

Can you integrate with my specific TMS?

We do not promise integration with any named product, because that depends on what your installation exposes, and that varies even between customers of the same vendor. We look at what your system accepts, then write by the most reliable route it offers. If the only way in is the screens, an agent can fill those, more slowly than an API.

What should I ask a TMS automation vendor before signing?

Ask what it writes and by which route, what happens to a document it cannot read, whether the original page stays linked to each field, and what leaves with you if you cancel. The answer to the third question matters most. A vendor who cannot explain the exception path has not built one.

Sources
Where to start

Does this cost you the same every month?

If the answer depends on who is handling it that week, the knowledge is sitting in a person rather than in the process, and it leaves when they do. Tell us which process behaves that way. One person reads it and one person replies, with what we would standardise first and why.

More on this: Automation, or all resources.

Tell us what breaks