What Is AI Dispatch? A Clear Definition

AI dispatch is software that works the freight dispatch loop for you: it finds loads across the sources you already use, ranks them by the money you actually keep after costs, drafts the outreach to brokers, and surfaces the handful of decisions that need a person. A human still approves every commitment. The term gets used loosely, so this is the clean version: AI dispatch is not a chatbot bolted onto a load board, and it is not a system that books freight behind your back. It is a tool that does the reading, math, and first-draft writing that eats a dispatcher's day, and then hands the judgment calls back to the dispatcher.

That distinction matters because "AI" in freight has been stretched to cover everything from a search filter to a marketing slogan. To pin it down, it helps to look at the actual work dispatch involves, where the time goes, and what software can and cannot take off a person's plate.

The dispatch loop, and why it is hard to automate

Dispatching a truck is a loop that repeats every time a load delivers. Find candidate loads. Figure out which ones pay after you subtract fuel, the miles you run empty to reach the pickup, tolls, and the driver's time. Contact the broker on the best option. Negotiate a rate. Confirm the details. Track the load to delivery. Then start over before the truck sits empty. Most of this is reading and comparison, not strategy — but the volume is brutal. A load board can show hundreds of postings on a single lane, each with a different rate, distance, and pickup window, and the good ones disappear in minutes.

The economics are what make the loop unforgiving. The American Transportation Research Institute put the marginal cost of operating a truck at roughly $2.26 per mile in its 2025 report (covering 2024 data), so a few cents per mile in either direction is the difference between a profitable week and a break-even one. Deadhead — the empty miles a truck runs to reposition — commonly eats 15 to 30 percent of total miles, and every one of those miles burns the same fuel while earning nothing. A dispatcher who eyeballs the loaded rate without subtracting deadhead and fuel can book a load that looks fine and loses money. Doing that math correctly, on every option, under time pressure, is exactly the kind of work software is good at and humans get tired of.

The market makes it worse. There are roughly 787,000 active carriers in the US (FMCSA, December 2023), and about 91.5 percent of them run ten trucks or fewer (ATA, 2025). Those small carriers compete against each other and against the roughly 27,000 freight brokers who sit between them and the shippers, taking a margin that averaged around 13.5 percent (DAT, 2023). A small carrier rarely has a full dispatch team; often it is the owner doing it at night. That is the gap AI dispatch is built to close — not by replacing the dispatcher, but by giving one person the throughput of several.

What AI dispatch actually does

Strip away the marketing and AI dispatch does four concrete things. First, it aggregates supply: it pulls postings from the load boards and broker sources you use, continuously, so you are not refreshing a board and re-typing the same lane search every few minutes. Second, it ranks. This is the part that earns the name — instead of sorting by posted rate, it estimates the all-in economics of each load: rate minus deadhead to the pickup, minus fuel at current prices, against the cost-per-mile floor you set. A $2.40 load 200 miles away can rank below a $2.10 load at your door, and the ranking says so. Third, it drafts outreach: given a load you want, it writes the email or message to the broker — the inquiry, the rate counter, the follow-up — in seconds, so the bottleneck is your approval rather than your typing. Fourth, it surfaces exceptions: the detention that is about to start billing, the rate confirmation that does not match what was agreed, the load that has gone quiet. It pushes the things that need a human to the top instead of letting them hide in a thread.

What it does not do is also part of the definition. It does not commit you to freight on its own. A person reviews and approves before anything is booked or any rate is accepted. It does not place autonomous voice calls negotiating with brokers — at least, the honest tools do not pretend to. In practice today the negotiation that AI handles well is the written kind: drafting and structuring email back-and-forth with brokers, where there is time to check the numbers and the draft sits in front of a human before it sends. Anyone selling fully autonomous AI that calls brokers and closes deals without a person in the loop is describing a product that does not match how this work safely runs.

The reason for keeping a person on the commitments is not caution for its own sake. Freight has real downside. Detention costs the industry an estimated $1.1 to $1.3 billion a year, double-brokering is on the rise, and cargo theft hit roughly $725 million in 2025 by CargoNet's count. A booking is a contract; a wrong one is money lost or a load stolen. AI is good at flagging the suspicious rate con or the broker that does not check out — and bad at being the one who is accountable for the decision. So it flags, and the dispatcher decides.

How it differs from a TMS or a load board

The fastest way to understand AI dispatch is to put it next to the two tools every carrier already touches: the load board and the transportation management system. They solve different problems, and AI dispatch solves a third.

A load board is a place to look. It shows you what is posted, but it does not tell you which load makes money for your truck from your current location, and it does not contact anyone. The sorting and the math are still yours. A TMS is a place to record. It is excellent at managing freight you have already booked — tracking, paperwork, settlements, compliance — but it starts working after the dispatch decision is made. Ask a TMS to find you a better-paying load right now and it has nothing to say; that is not its job.

AI dispatch sits in the gap between them: the work of turning a board full of postings into a short list of loads worth a call, and turning that into outreach that is ready to send. It does not replace either tool. It reads from the load board and it can feed the TMS once a load is booked. It replaces the manual labor in the middle — the part a person currently does by switching between tabs, copying rates into a spreadsheet, and typing the same broker email for the hundredth time.

Who it is for, and who it is not

AI dispatch helps most where the gap between the work and the people doing it is widest. That is the small-to-mid carrier and the owner-operator. When 91.5 percent of carriers run ten trucks or fewer, the typical operation does not have a bench of dispatchers; it has one or two people, or the owner, covering sourcing, negotiation, tracking, and paperwork at once. For them, the value is not exotic — it is getting the all-in math right on every load instead of guessing, and reclaiming the hours spent refreshing boards and retyping emails. A dispatcher's median pay was about $46,860 (BLS, 2023); for a small carrier, the choice is often between adding a person at that cost and giving the person they have more reach. AI dispatch is the second option.

It is genuinely less useful in a few cases, and it is worth being honest about them. A carrier running entirely on dedicated, contracted freight — same lanes, same shippers, locked rates — has little to source and little to negotiate, so the ranking and outreach engine has nothing to chew on. A large fleet with a mature, well-staffed dispatch operation and deep direct-shipper relationships already has much of this throughput; AI helps at the margins there rather than transforming the work. The sweet spot is the carrier living in the spot market, where loads turn over constantly, rates move, and the difference between a good book and a bad one comes down to doing fast, correct math on a lot of options — over and over, all day.

The broader trend is real but easy to overstate. Surveys put AI adoption across supply chains high — Gartner has reported around 67 percent and ABI around 94 percent — but those numbers cover everything from warehouse robotics to demand forecasting, not dispatch specifically. The honest read is that the tooling for dispatch has matured to where a one-truck operator can use it, not that the industry has fully adopted it. Adoption at the carrier level is early, which is precisely why understanding what the category actually is — and is not — is worth the time.

The takeaway

AI dispatch is software that works the dispatch loop: it aggregates loads, ranks them by what you keep after deadhead and fuel, drafts the broker outreach, and pushes the exceptions that need a person to the front — while a human approves every commitment. It is not a TMS, which records freight you have already booked, and it is not a load board, which only lists. It is the working layer between them, and it earns its keep by getting the economics right at a speed and consistency a tired person can't match, on the spot-market freight where that math decides the week.

If you want to see what that looks like in the place you already work, Numeo Spot runs as a Chrome extension inside DAT and Truckstop, scoring each posting on all-in economics so the ranking is in front of you without leaving the board.