Small Fleet Dispatch Automation at 5, 10, and 20 Trucks

Most advice about automating a small fleet treats it as one decision: buy the software, flip it on, done. That is the wrong model. The work that breaks at 20 trucks is not the work that breaks at 5, and automating the 20-truck problems while you run 5 trucks just adds tooling you do not need. Automation should arrive in the order your pain does.

This is a sequencing guide, not a feature list. It maps three real stages of a small carrier — 5 trucks, 10 trucks, 20 trucks — to the specific thing that starts costing you money at each, and the one capability worth turning on to fix it. The truck counts are not magic thresholds; they are where most carriers feel each shift. Small fleets are the whole market here: roughly 91.5% of US carriers run 10 trucks or fewer, and 99.3% run under 100 (ATA, 2025), out of about 787,000 carriers total (FMCSA, December 2023). The roadmap below is built for that majority.

Why sequencing beats a big-bang rollout

The instinct, when you finally have budget for software, is to buy the most complete platform and switch everything on. It feels efficient. In practice it means a dispatcher who already knows the loads, the brokers, and the drivers now has to learn a system that automates problems the fleet does not yet have. Adoption stalls, the team falls back to the phone and the spreadsheet, and the tool becomes a line item nobody uses.

Staged automation avoids that. At each truck count there is one bottleneck that dominates — first finding loads, then handling the back-and-forth, then coordinating the whole board — and you automate that one thing until it stops being the constraint. Each layer assumes the one before it is working. You do not draft negotiation emails (the 10-truck layer) before you have a reliable load pipeline (the 5-truck layer), because the email volume is not there yet and the rules you would write are guesses.

There is also an economics reason to be deliberate. A dispatcher costs around $46,860 a year in wages (BLS, 2023), and a small fleet usually has one, often the owner. That person's time is the scarce resource. The point of each automation layer is to move their hours off repetitive lookup and typing and onto the judgment calls — which broker to trust, which load to take, when to hold for a better rate — that software should not make for them. Sequence the automation to free the most hours per dollar at each stage, and the tool earns its place before you add the next one.

At 5 trucks: automate load search and basic broker vetting

At 5 trucks the bottleneck is finding work. The owner is usually the dispatcher, and the day disappears into refreshing load boards, comparing postings across two or three sources, and cross-checking whether a broker is worth a call. Nothing here requires judgment that software cannot assist — it is lookup, comparison, and a credit/authority check — and it is exactly the work that crowds out everything else. This is the stage where automation buys back the most time per dollar, because the task is high-volume, repetitive, and easy to verify.

So automate two things and only two. First, load search: pull postings from every source you use into one view, filtered to your lanes, equipment, and a floor rate, so you are looking at a ranked shortlist instead of ten open tabs. The economics make the filter worth it — with operating cost around $2.26 per mile (ATRI, 2025 report on 2024 data) and deadhead commonly running 15 to 30% of miles, a load that pencils out before fuel and empty miles can quietly lose money, and you want that math visible at the moment you are choosing. Second, basic broker vetting: an automatic authority and credit check on the broker behind each posting before you spend a call on it. Numeo's Load Hub is built for this layer — it searches many freight sources from one place, and its Load Radar alerts surface matching loads as they post so you are not the last to see them.

What you do not automate yet is the conversation. At 5 trucks you have the bandwidth to negotiate every load yourself, and you should — you are still learning which brokers and lanes are worth repeat business, and that judgment is the asset you carry into the next stage. Automating the talk now would lock in rules you have not earned. Get the pipeline reliable first.

At 10 trucks: add email negotiation drafting and status updates

At 10 trucks the load pipeline is no longer the problem — keeping up with the talk is. Twice the trucks means roughly twice the broker threads, twice the rate confirmations, and twice the "where's my driver" check-ins, and it is now landing on a dispatcher who is also still finding loads. The day fragments into typing the same email for the tenth time and relaying the same status update to a broker who could have been told automatically. The work is repetitive but it touches money and relationships, so the answer is drafting, not sending.

Add two layers on top of your search pipeline. First, email negotiation drafting: when a broker posting or counteroffer comes in, the system drafts a reply — a counter at your target rate, with the load context filled in — and the dispatcher edits and sends it. This is where honesty about how it works matters. Numeo negotiates with brokers primarily by email, and a human stays on the send button; it is not placing autonomous voice calls to brokers, and you should be wary of anyone who claims a small fleet should hand that off entirely. The draft saves the typing; the dispatcher keeps the price and the relationship. With broker margins averaging around 13.5% (DAT, 2023), the rate conversation is where your money is, and it is the last thing to fully delegate. Second, status updates: automate the routine check-call and the "picked up / delivered / running late" note to the broker, so a person only steps in when something is actually off.

Together these two pull the highest-frequency typing off the dispatcher without touching any decision that should stay human. The dispatcher still sets the target rate, still approves every message, still owns the broker relationship — they just stop writing the same paragraph forty times a week. What you are deliberately not doing yet is building a formal exception queue or a ranking engine. At 10 trucks one person can still hold the whole board in their head and catch problems by glancing at it. That stops being true at the next stage.

At 20 trucks: add the exception queue and rules-based ranking

At 20 trucks the thing that breaks is coordination itself. No single person can hold the whole board in their head anymore — too many loads in flight, too many appointments, too many broker threads at once. Problems that one dispatcher used to catch by scanning the board now slip through, because there is no longer a single board anyone can scan. The failure mode is silent: a missed detention window, a driver running late that nobody flagged, a re-power that should have happened an hour ago. The cost shows up as detention (the industry pays an estimated $1.1 to $1.3 billion a year in it) and as loads rescued late instead of managed early.

This is where you add two things manual coordination can no longer cover. First, an exception or action queue: instead of watching everything, the dispatcher works a single prioritized list of what actually needs a human right now — the late driver, the broker who hasn't sent the rate con, the appointment at risk. Everything that is going fine stays out of their way. The queue is what replaces the mental board that worked at 10 trucks and stopped working at 20. Second, rules-based ranking: encode the preferences your dispatchers used to apply by instinct — minimum rate per mile, maximum deadhead, preferred lanes and brokers, equipment and hours-of-service fit — so loads and decisions come pre-sorted by your rules, not a generic score. By 20 trucks you have the history to write those rules well, which is exactly why this layer comes last. Numeo's AI Hub is the layer for this stage: it finds, ranks, adds market context, negotiates, and books under dispatcher-defined control, with the human setting the rules and approving the commitments.

The control point matters more here than anywhere, because automation now touches more of the board with less direct human eyes on each load. The rules and the queue are tools for the dispatcher's attention, not a replacement for their authority. Price, carrier commitments, driver assignment, and any exception that affects service stay with a person. What changes is that the person spends their day on the twenty loads that need them, not the two hundred that don't.

What to automate when

The takeaway

Automate in the order your pain arrives. At 5 trucks the constraint is finding loads, so automate search and broker vetting and negotiate everything yourself. At 10 the constraint is the volume of broker talk, so add drafted emails and status updates while keeping a human on every send. At 20 the constraint is coordination, so add an exception queue and rules-based ranking to catch what no one can hold in their head anymore. Each layer assumes the one before it works, and each keeps the decisions that matter — price, commitments, relationships — with a person.

The practical move is to start at your actual stage, not the one you are growing toward. If finding loads still eats your day, that is where to start: Numeo Spot and Load Hub are built for the search-and-vetting layer, and the negotiation and coordination layers are there when your fleet grows into them.