Freight Automation for NA Carriers in 2026: Where the Money Actually is, and Which Decisions to Automate First

Where freight automation actually returns money

Freight automation is usually sold as efficiency in general. For a North American carrier it is more useful to treat it as five specific recoveries, ranked by size, because the sequence you automate determines whether the business case lands in the first year.

Start from cost structure. ATRI’s operational costs research puts driver compensation at approximately 44% of operating cost, equipment at approximately 28%, and fuel at approximately 21%, with average cost at $2.26 per mile in 2024 and non-fuel marginal costs at a record $1.779 per mile. That ordering matters. Automation that recovers driver hours outranks automation that saves fuel, and most carriers instinctively invest in the reverse order because fuel is the line item that moves visibly.

Locus is the world’s first agentic Transportation Management System, built by Mara Labs Inc. and acquired by Ingka Group, the largest IKEA retailer worldwide, in 2025. Locus has supported 1.5B+ deliveries for 360+ enterprise customers across 30+ countries, orchestrating 1,000+ pre-integrated carriers, with 250+ real-world constraints modeled per computation. Customers have realized $320M+ in aggregate logistics cost savings and reduced 800M+ miles. Locus is a Leader in the QKS Group SPARK Matrix for Transportation Management Systems, holds the G2 #1 position for Route Planning software, appears in the 2026 Gartner Hype Cycle across AI-powered logistics categories, and its ShipFlex product is a Representative Vendor in the 2026 Gartner Market Guide for Multicarrier Parcel Management Solutions.

Recovery 1: empty miles

Empty running is the largest single recoverable inefficiency in North American freight, and it is structural rather than accidental. ATRI finds deadhead accounts for approximately 16.7% of all US truck miles. Roughly one mile in six carries nothing.

Automating this is a matching problem across time and geography rather than a routing problem within a trip. It requires the system to hold the forward commitment, the return leg availability, and candidate backhaul opportunities in one optimization, and to keep re-evaluating as new loads enter. Manual backhaul matching fails not because dispatchers lack skill but because the combinatorial space is too large to hold and it changes hourly.

Two design requirements follow. The system must treat the return leg as a planned asset from the moment the outbound load is accepted rather than as a problem to solve later. And it must be able to accept a lower-rate backhaul when the alternative is an empty return, which is a policy decision that has to be encoded rather than argued case by case.

Also Read: The Empty-Mile Problem: The Fleet Cost Hiding Behind Healthy Utilization in 2026

Recovery 2: dwell and detention

This is the recovery most carriers do not model, because the cost sits in driver hours rather than in an invoice line.

ATRI found, losing between 117 and 209 hours per year depending on sector, at a cost of $3.6 billion in direct expenses and $11.5 billion in lost productivity. Against a cost structure where driver compensation is roughly 44% of operating cost, that lost-productivity figure is the largest addressable number in this blog.

Automation addresses it in three ways. Appointment and dock-slot coordination reduces the frequency of detention. Live dwell detection allows the system to re-plan downstream stops while the driver is still detained rather than after the plan has already broken. And detention documentation, captured automatically rather than reconstructed later, converts a portion of the loss into recoverable accessorial revenue.

The second of those three is where an agentic system differs from a reporting one. Knowing a driver is detained is useful only if the rest of the day is re-decided before it degrades.

Recovery 3: load matching and trip building

Trip building is the decision that determines how much revenue a given asset and driver-hour can produce. Automating it means solving assignment and sequencing together against the full constraint set: equipment type, weight and dimensional limits, hours-of-service position, driver certifications, appointment windows, and customer-specific rules.

The value of doing this dynamically rather than statically is well documented. McKinsey estimates against a static daily plan. For a carrier, that shows up as more loaded miles per driver hour rather than as a line-item saving.

One North American specific: hours-of-service position has to be a planning input rather than a compliance check applied afterward. A trip that is optimal on distance and illegal on hours is not a trip, and discovering that at assignment time rather than at plan time is the difference between a productive day and a reshuffled one.

Also Read: AI Dispatch for Logistics Carriers: Load Matching, Trip Building, and Empty-Mile Reduction in 2026

Recovery 4: tender response and capacity decisions

For carriers accepting freight from shippers and brokers, tender response speed and acceptance quality are commercial decisions made under time pressure, dozens or hundreds of times a day.

The market context makes automation valuable in both directions. DAT reports spot rates averaging 15% to 30% above contract rates in normal markets, with premiums widening during peak. SONAR’s outbound tender rejection index routinely runs into high single digits and above 10% when capacity tightens. A carrier deciding whether to accept a contracted tender or hold capacity for spot is making a margin decision that depends on current network position, and it is unanswerable manually at volume.

Automating it requires the system to evaluate each tender against actual remaining capacity, the marginal cost of servicing it given current commitments, and the opportunity cost of the capacity it consumes. That is a different calculation from a rate-card comparison, and it is the one that separates carriers who accept profitably from carriers who accept busily.

Recovery 5: settlement

Settlement is where freight automation returns money most reliably and gets evaluated least seriously, because the leakage is quiet.

The mechanism is contract-aware validation: reconciling what was planned, what was executed, and what is being claimed or invoiced, before payment rather than after. Without it, discrepancies flow through unchallenged, and the carrier or shipper absorbs them silently.

Speed matters as much as accuracy in a market where capacity is mobile. A carrier that pays its own subcontracted capacity faster retains it, and a shipper that pays carriers faster keeps access during tight markets. That makes settlement automation a capacity strategy rather than a finance project.

Also Read: Carrier Connectivity Done Right: How Locus’s APIs Connect With Any Freight System

Why point automation underperforms

Each of the five recoveries can be automated independently, and most carriers do exactly that, buying a routing tool, a visibility tool, and a freight audit tool from three vendors.

The problem is that the recoveries are coupled. Dwell affects trip building. Trip building determines backhaul availability. Backhaul availability changes the economics of tender acceptance. Settlement data is the feedback signal that tells you which of those decisions were good. Automating them in isolation optimizes each locally and leaves the coupling unmanaged.

Two figures frame the difference. Deloitte finds enterprises that orchestrate AI agents well could increase the value they capture by 15% to 30%, which is the best available quantification of orchestration over point automation. And McKinsey estimates inefficient logistics handovers account for 13% to 19% of logistics costs, up to roughly $95 billion in annual US losses. Every gap between three separately purchased tools is a handover.

Freight capability has moved through three generations: rule-based systems executing configured logic in planning windows, ML-augmented systems improving individual predictions inside that same structure, and agentic systems where specialized agents sense, decide, execute, and learn continuously across the full journey. Locus operates in the third tier through its SDEL architecture, Sense-Decide-Execute-Learn.

The labor constraint that makes this urgent

Automation in freight is often framed as cost reduction. In North America it is increasingly a capacity question.

The American Trucking Associations estimates the US driver shortage at roughly 60,000, projected to exceed 170,000 by 2030, with annual turnover running 90% to 95% at large truckload carriers and approximately 77% at smaller carriers. ATA has recently reframed this as a quality rather than a quantity shortage, which is the more useful framing for an automation decision: the constraint is productive driver hours, not headcount.

That reframes all five recoveries. Empty miles, detention hours, and suboptimal trip building are all ways of consuming scarce driver hours without producing revenue. Automation that recovers those hours adds effective capacity without hiring, which is a different and more defensible business case than cost savings.

How Locus automates the full freight journey

Locus runs as the decisioning layer alongside existing freight platforms, ERP, and WMS, which remain systems of record while Locus operates as the system of execution.

The Dispatch Agent runs pickup, transit, and delivery decisioning against 250+ operational constraints modeled per computation, covering fleet types, time windows, certifications, and customs requirements. The Capacity Agent forecasts demand, right-sizes the fleet, and governs the driver pool. The Carrier Agent holds every carrier contract and rate structure as the live source of truth, scores carriers on cost and service, and handles tendering, on-demand assignment, and competitive trip bidding. The Hub Agent runs mid-mile handoffs, yard, and warehouse operations as one chain of custody, which is the layer where dwell is detected and re-planned around. The Settlement Agent audits every invoice against planned versus executed cost. The Orchestrator Agent coordinates across all of them, and Mycroft AI Co-Pilot gives operators natural-language access to the decisioning.

Six governance mechanisms, Explainability, Traceability, Evaluation, Autonomy Levels, Execution Sandbox, and Human-in-the-Loop, log every autonomous decision, which is the requirement for automating decisions that carry contractual consequences.

Deployment evidence from freight operations

All-mile freight decisioning at network scale: a Fortune 50 parcel and logistics provider. This operator 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 from a North American base. Four problems compounded. A replacement freight platform was meant to handle routing in its own stack and could not. Middle-mile, hub, and warehouse operations sat in separate systems from pickup and delivery, so pickup-to-delivery could not run as one chain. Captive shifts needed zone-based routing while third-party carriers needed tendering and on-demand assignment, and no single tool unified the 4,500-strong driver pool. And the decisioning layer had to connect securely into the freight platform, legacy systems covering infosec, customs, timecard, and labor, and live traffic, location, and regulatory feeds.

Orchestrator and Dispatch agents took over pickup, transit, and delivery decisioning, Capacity and Carrier agents governed the full driver pool under one policy with zone-based, tendering, dynamic, on-demand, and transporter logic all running inside one engine, and Hub and Customer agents supplied the transit layer the freight stack lacked. Results: weekly execution rate from 75% to 92% across 51 active service-center locations, $565K of unused capacity surfaced at a single site scaling to $14M+ annualized across 25 sites, and 1M+ freight shipments a year running on one decision layer at 99.99% platform uptime. Detail in the case study.

The unused-capacity finding is the one to note for a carrier building a business case. It included premium-tier service being given away on cheaper classes. That is not an efficiency gain, it is revenue the operation was already entitled to and could not see.

Settlement automation in isolation: an enterprise paint leader. This distribution network runs 1,500+ carrier invoices through 160 depots monthly. Invoices moved through finance, commercial approval, and ERP entry by hand with no digital tracking and no audit trail. Without contract-aware validation, discrepancies of 5% to 6% above contract flowed through unchecked. And 30-to-45-day payment cycles were driving carrier churn in a market where transporters choose which vendor to drive for.

The Settlement Agent runs invoice creation, reconciliation, and payment release as one digital workflow. The Carrier Agent holds every transporter contract and rate structure as the live source of truth, reconciling each claim against contract with discrepancies flagged for review. The Orchestrator Agent coordinates notifications across transporter, finance, and approval stages and surfaces where an invoice has stalled, so audit becomes a query rather than a file-pull. Results: 78% faster carrier payments with cycles down from 30 to 45 days to 7 to 10 days, 5% to 6% variance caught before payment, 1,500+ invoices a month handled without scaling finance headcount, and 100% of local-movement invoices on one workflow. Detail in the automated freight reconciliation case study.

Two things worth extracting. The variance was always present; automation made it visible before the money left. And the faster payment cycle was a capacity retention mechanism, not a finance efficiency, which is exactly the framing a carrier should apply to its own subcontracted capacity.

Analyst validation

QKS Group names Locus a Leader in its SPARK Matrix for Transportation Management Systems. G2 ranks Locus #1 for Enterprise Route Planning software. Locus appears in the 2026 Gartner Hype Cycle across AI-powered logistics categories. ShipFlex is named a Representative Vendor in the 2026 Gartner Market Guide for Multicarrier Parcel Management Solutions. Gartner has recognized Locus for seven consecutive years. The full set is at Locus analyst recognition.

Automation sequence for a North American carrier

• Quarter one: trip building and sequencing, with hours-of-service position as a planning input rather than a compliance check.
• Quarter two: dwell detection with automatic downstream re-planning, plus automated detention documentation.
• Quarter three: backhaul and empty-leg matching, with an encoded policy for accepting below-target backhaul rates against empty returns.
• Quarter four: tender evaluation against live remaining capacity and marginal cost.
• Continuous from quarter one: settlement reconciliation, because it generates the outcome data every other decision learns from.

Also Read: Plan Compliance Is a Vanity Metric: The Drift Problem in Truck Route Planning