# No More Empty Runs: How an AI Agent Plans a Transport Company's Routes and Loads in Real Time

> An AI agent for freight routing consolidates loads and reacts to change in real time. How growth companies cut empty runs and raise fleet utilization.

- Published: 2026-09-19
- Author: Master Mind
- Canonical: https://aimaster.fi/en/artikkelit/tyhjana-ei-ajeta-nain-ai-agentti-suunnittelee-kuljetusyrityksen-reitit-ja-kuorma

An empty kilometer costs a transport company exactly as much as a full one — without generating a single euro of revenue. Yet much of Finland's road freight still runs on a route plan drawn up once in the morning that never adapts to the day. The fix isn't more manual work for dispatchers. It's an AI agent for freight routing and load consolidation.

A dispatcher opens their planning tool in the morning, builds routes and loads with the best information available — and then the day changes. A new pickup comes in. A customer shifts a delivery window. A road gets blocked. The static plan doesn't react, and the chance to combine two loads on the same route is lost. The result shows up directly in margin: extra kilometers, overtime, and drivers waiting with an empty vehicle for the next instruction.

## What is an AI agent for freight routing and load consolidation?

An AI agent for freight routing is a system that continuously tracks data from vehicles, orders, and traffic conditions and makes independent decisions about routes and load combinations. Unlike traditional routing software that calculates one plan once in the morning, the agent updates the plan throughout the day and reacts to new orders, cancellations, and congestion without forcing the dispatcher to recalculate everything by hand.

## How does a routing agent work in practice?

A working solution isn't one all-knowing program — it's a set of smaller agents that share responsibility. Each agent handles its own task and passes information forward when conditions change.

- A dispatch agent keeps the overall picture of all open orders and vehicles, and proposes the best combination whenever a new pickup enters the system.
- A vehicle agent represents a single truck: it knows its location, remaining capacity, and the driver's remaining hours in real time.
- An order agent tracks one delivery's time window and raises a flag if the schedule risks slipping due to traffic or a delayed pickup.

When these agents exchange information in a standardized way, they can negotiate a new solution in seconds instead of a dispatcher working through the options by hand. How agents from different vendors actually speak a common language is covered in more detail in the article [A2A protocol](https://aimaster.fi/en/artikkelit/a2a-protokolla-nain-eri-valmistajien-ai-agentit-oppivat-puhumaan-toisilleen).

## Static routing vs. an AI agent — what's the difference?

| Feature | Static routing | AI agent |
| --- | --- | --- |
| Plan updates | Once in the morning | Continuously, all day |
| Reaction to a new order | Next day's plan | Seconds, mid-route |
| Load consolidation | Relies on dispatcher memory | Calculated for every open order |
| Data source | Morning order data | GPS, telematics, TMS and ERP in real time |
| Who decides | Human calculates manually | Agent proposes, human confirms exceptions |

## What data does a routing agent need to function?

An agent is only as good as the real-time picture it has access to. It needs a connection to at least four sources: order data from the transport management system (TMS), billing and customer data from the ERP, vehicle telematics, and location data from driver apps. Many transport companies run older systems that don't talk to each other directly — the fix is to build an [API wrapper on top of the legacy system](https://aimaster.fi/en/artikkelit/legacy-jarjestelman-avaaminen-ai-agenteille-nain-kasvuyritys-rakentaa-api-kaaree) that safely opens the data to the agent without replacing the whole system.

Building this data foundation is exactly what [Master Layer](https://aimaster.fi/tuotteet/master-layer) does: it connects a company's existing systems safely for AI use without requiring a TMS or ERP replacement. Only once data flows reliably between systems can the agent make decisions a dispatcher can actually trust.

## How does a growth company adopt a routing agent without a big-bang IT project?

Adoption doesn't require replacing everything at once. The first step is mapping out where routing and load consolidation are losing the most money right now — a phase productized as [Master Plan](https://aimaster.fi/tuotteet/master-plan). Data is then connected through Master Layer, and finally [Master Mind](https://aimaster.fi/tuotteet/master-mind) handles the actual routing and consolidation decisions as part of daily operations. Development proceeds in 3-day sprints, so the first measurable impact on route planning appears before a traditional consulting project would have finished its mapping phase.

The dispatcher's role doesn't disappear. The agent proposes and handles routine calculations, but exceptions — a difficult customer, a special shipment, a vehicle breakdown — always go to a human to resolve. The goal is to cut manual calculation, not to remove the dispatcher's judgment where it's genuinely needed.

## Where should you start if routing is still fully manual?

Start by measuring the current state: how many vehicles return empty, and how often same-day orders go unconsolidated on the same route. These two numbers reveal whether a routing agent creates value fastest by raising utilization or by cutting empty runs. Do this mapping before choosing a system, not after.

## Frequently asked questions

### What does an AI agent for freight routing cost?

Cost depends on scope and how many systems the agent needs to integrate with. The sprint model makes cost predictable: development proceeds in 3-day cycles, billed per completed sprint. The first step is mapping where routing loses the most money — that determines the budget, not the other way around.

### Does an AI agent replace the dispatcher?

No. The agent handles routine calculations and proposes routes, but exceptions — a vehicle breakdown, a special shipment, a difficult customer situation — are still resolved by a human. The goal is to cut manual calculation, not remove dispatcher judgment.

### Does a routing agent work with a smaller fleet?

Yes. The benefit scales down with fleet size, but even a few dozen vehicles generate enough combinations worth calculating automatically instead of a dispatcher doing it by hand every morning.

### How does the agent integrate with an existing TMS or ERP?

The agent needs a real-time connection to order, billing, and telematics data. If systems are older and lack a ready interface, a lightweight API wrapper is built on top that safely opens the data without replacing the whole system.

### How long does adoption take?

First measurable results typically appear within a few weeks, since development proceeds in 3-day sprints rather than a months-long planning phase. Scope depends on how many systems need to be integrated.