TMS Requirements for USMCA Cross-Border Freight: What Breaks at the Border

The flow most TMS deployments treat as domestic

North American cross-border freight is not an edge case sitting at the margin of a domestic network. It is a primary flow.

The US Bureau of Transportation Statistics reports USMCA transborder freight totalling 1.6 trillion dollars in 2024, with US-Canada flows at 761.2 billion dollars and US-Mexico at 839.9 billion dollars. Trucking carried 55.5 percent of flows with Canada and 72.5 percent with Mexico, and surface transportation accounted for 77.1 percent of all North American freight flows by value.

Most TMS implementations serving those flows were configured for domestic movement and extended outward. A cross-border lane becomes a domestic lane with a longer transit time and, if the operation is careful, a wider delivery window. That configuration works until something goes wrong at the border, at which point the system has no representation of what is happening or why.

The cost of that gap is a handover cost, and handovers are expensive in aggregate. McKinsey estimates that inefficient logistics handovers account for 13 to 19 percent of logistics costs, as much as 95 billion dollars annually in the US alone. A border crossing concentrates several handovers into one point in the journey.

Four things that break at the border

1. Status continuity

A domestic shipment produces a continuous status stream from the carrier. A cross-border shipment produces that stream, then stops, then resumes.

The gap is the clearance process, and it sits outside the carrier’s event model because the carrier is not the party performing it. From the TMS’s perspective the shipment has simply gone quiet, which is indistinguishable from a data feed failure. Operations teams learn to interpret the silence, which works until volume grows past the point where interpretation scales.

This is not a tracking problem to be solved with a better carrier feed. It is a modelling gap: the status taxonomy has no state representing “in clearance,” so there is nowhere for the information to go even when it exists.

2. Carrier handoff

Many cross-border movements involve more than one carrier, with a drayage or transfer operator handling the crossing itself and different carriers operating either side.

A TMS modelling a shipment as a single carrier assignment cannot represent this, so the movement is either split into unrelated records, which breaks end-to-end visibility, or forced into one record that misrepresents who holds the freight. Both produce the same downstream symptom, which is that nobody can answer where the shipment is without a phone call.

3. Documentation as a precondition

Domestic freight treats paperwork as something that accompanies or follows the movement. Cross-border freight treats it as a gate: incomplete or inconsistent documentation stops the truck, and correcting it at the border costs far more than correcting it before departure.

The planning implication is that documentation completeness has to be a dispatch precondition rather than a downstream record. A TMS that can dispatch a load whose paperwork is incomplete is a TMS that will occasionally dispatch a load that will sit at the border.

4. Dwell expectations calibrated on the wrong data

Border dwell varies by crossing, by time of day, by day of week, and by commodity, in ranges that domestic facility dwell does not approach. A model using a network-wide average will be wrong in both directions, over-allocating time on smooth lanes and under-allocating on congested ones.

Dwell is already the underestimated variable in domestic freight. ATRI found drivers were detained at 39.3 percent of all stops in 2023, losing 117 to 209 hours per year depending on sector, at a cost of 3.6 billion dollars in direct expenses and 11.5 billion in lost productivity. Border crossings add a category of dwell with wider variance and less operator control.

What a cross-border-capable TMS has to model

The first row carries the others. Without a modelled clearance state, there is nowhere to attach a duration model, an exception reason, or a third-party event.

Why this is an execution gap, not a visibility gap

It would be reasonable to read the above as a case for better visibility tooling. It is not, quite.

Knowing a shipment is held in clearance is necessary and does not by itself change any outcome. What changes outcomes is whether the operation can act on it: re-plan the downstream leg, reassign the receiving-side carrier whose appointment will now be missed, notify the consignee with a revised window, and reprioritise the loads behind it.

That is a dispatch capability rather than a reporting one, and it is where most stacks stop. Gartner found that while 95 percent of supply chains must react quickly to change, only 7 percent can execute decisions in real time. A cross-border delay is precisely the case where the gap between detection and action is expensive, because the downstream leg has its own appointment, its own carrier, and its own capacity that is about to be wasted.

What to ask vendors

Eight questions specific to cross-border capability.

1. Is customs or clearance a modelled state in your status taxonomy, or does the shipment simply have a gap?
2. Can a single shipment be represented as connected legs across multiple carriers, with delay on one propagating to the ETAs of the others?
3. Can dwell expectations be set per crossing and per direction, and can they learn from our history?
4. Can documentation completeness be enforced as a condition of dispatch?
5. Can you ingest status events from brokers and other non-carrier parties?
6. When a clearance delay occurs, what happens automatically to the downstream leg?
7. How is party and address data handled across jurisdictions, and what validation exists?
8. What in your platform is coupled to current duty or tariff treatment, and what would change if that treatment changed?

Question eight is worth pressing on. Trade policy has been volatile, and a platform with policy specifics hard-coded into workflows will need vendor work each time treatment changes. Architecture that treats duty and tariff as configurable data survives; architecture that treats it as logic does not.

Where Locus fits

Locus, the world’s first Decision-Intelligent, Agentic TMS, is built around multi-leg execution, which is the structural requirement cross-border imposes. A shipment is represented as connected legs rather than as separate records, so a delay on one leg recalculates the feasibility of the legs after it.

Within DiSCO, the Dispatch agent re-sequences on live events against 250+ real-world constraints, the Capacity agent identifies alternative capacity when a downstream leg is no longer serviceable, the Hub agent manages facility-side constraints and handoffs, and the Customer agent issues the revised commitment. Six governance mechanisms bound autonomous action, including autonomy levels, which matters where a re-plan has customs or contractual consequences and should stay recommend-only until confidence is established.

Locus has been recognized by Gartner for seven consecutive years, featured in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, named a Leader in TMS by QKS Group (SPARK Matrix), and ranked #1 in Route Planning on G2’s 2026 Best Software Awards. In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently.

Two North American deployments show the relevant mechanisms. A Fortune 50 parcel and logistics provider runs one of the world’s largest multimodal freight forwarding operations across air, ocean, and ground, moving 1M+ freight shipments a year across a 120-country footprint. Locus connects into the freight platform and legacy systems including customs, timecard, and labour, alongside live traffic, location, and regulatory feeds, governing 4,500+ drivers under one policy. Weekly execution moved from 75 percent to 92 percent across 51 service-center locations, with every autonomous decision logged for explainability, traceability, and human override.

A leading North American retailer moves freight across ocean, rail, and road into a multi-hundred-store network, previously on six disconnected systems where planning ran leg by leg and exceptions surfaced only after delays had reached store service. Consolidating onto one decision layer, with the Hub agent orchestrating DC, yard, and ocean and rail transit, produced 99 percent-plus on-time store delivery, exceptions resolved in under two hours, and 1 million dollars-plus in savings with break-even inside the first year.

Where to start

Before evaluating platforms, run one diagnostic on your own data.

Take your last quarter of cross-border shipments and measure the gap between the last carrier status before the border and the first status after it. That interval is the period your current system has no representation of. Then compare its variance to your domestic facility dwell variance.

If the border interval is both long and highly variable, and your ETA model treats it as a constant, you have quantified the problem in terms your operations team and your CFO will both recognise. That number is also the baseline against which any platform change should be measured.