Ask why a driver sat for four hours and the answer is almost never "the facility had no capacity." It is that the dock ran behind and nobody told the driver. The appointment was booked for a window that stopped being realistic that morning. The paperwork was not ready and nobody checked. The load was not staged and nobody said so.

The cost of this is not ambiguous. The American Transportation Research Institute puts driver detention at roughly $15.1 billion a year — about $3.6 billion in direct expenses and $11.5 billion in lost productivity. Drivers are detained at 39.3% of all stops and accumulate 117 to 209 hours of waiting annually. For individual drivers it means an earnings reduction of $1,281 to $1,534 per year.

Detention typically bills at $50 to $100 per hour after a two-hour free window. Billing it recovers some money. Preventing it is worth more, and it is achievable, because the underlying failure is informational.

The three moments where detention becomes preventable

Detention has a timeline, and it has three intervention points. Each is earlier and cheaper than the last.

• 24-48 hrs out — What is knowable: Appointment vs. current facility schedule · What prevention looks like: Rebook before the driver dispatches
• 2-6 hrs out — What is knowable: Live ETA vs. booked window · What prevention looks like: Adjust arrival or reschedule
• At arrival — What is knowable: Dock position and expected wait · What prevention looks like: Tell the driver; offer alternatives

The first two are the valuable ones and the ones nobody works, because working them requires somebody to re-examine bookings that were already made. Nothing prompts that. A confirmed appointment feels settled, and the drift that invalidates it happens quietly.

The third does not prevent the wait, but it changes the experience substantially. A driver who knows they are third in line and looking at ninety minutes can eat, rest, or take hours-of-service decisions deliberately. A driver who has no information sits, and remembers.

Why the appointment stops matching reality

An appointment is a prediction made days in advance by two parties who each have their own reasons for it to slip.

On the facility side, the schedule compresses as earlier appointments run long, labour availability changes, or a receiving priority shifts. On the carrier side, the ETA moves with traffic, prior stops and hours-of-service limits.

Both sides know their own drift. Neither routinely tells the other, because doing so means someone noticing and initiating a conversation about a booking that was already agreed. That is exactly the persistent, low-status, easy-to-defer work that gets skipped every day — which is what makes it a good automation candidate rather than a discipline problem.

Can AI actually reduce driver detention?

Yes — by continuously comparing live carrier ETA against the booked appointment window and initiating a rebooking conversation as soon as the two diverge, rather than discovering the mismatch when the driver arrives. Most detention is a coordination failure that becomes irreversible only at the gate.

The operating loop is straightforward:

1. Monitor continuously. Every confirmed appointment re-checked against live ETA, not once but throughout.
2. Detect divergence early, using a threshold that leaves time to act.
3. Open the conversation immediately with facility and carrier, in parallel, across whichever channels each actually responds on.
4. Confirm and redistribute the new window to everyone, including the driver.
5. Escalate when no alternative works or the load is time-critical.

Step one is the part that essentially never happens manually. Nobody re-verifies yesterday's confirmed appointments against today's ETAs, because there is no trigger and no time.

The carrier relationship dimension

There is a commercial argument here that is easy to miss when detention is treated as a billing line.

With 86% of brokers reporting that capacity is harder to find and 72% expecting further tightening, carriers are increasingly choosing who they work with. Detention is one of the most concrete reasons a carrier declines your freight — not because of the money, but because a driver's day gets destroyed and the dispatcher's plan with it.

A broker who reliably keeps drivers informed and reschedules before a wait happens becomes a preferred customer. That preference converts into coverage on the lanes where coverage is hard, which is worth considerably more than the detention charges you avoided.

The same applies to facilities. A consignee that develops a reputation for holding drivers finds carriers pricing that in — which becomes the shipper's cost, expressed as a higher rate rather than as a detention line.

When it does happen, capture it properly

Prevention does not reach 100%, and the residual has to be billable.

That depends entirely on evidence: arrival and departure timestamps, and something authoritative establishing both. If those exist only as a driver's phone call to dispatch, the charge is unprovable three weeks later and gets written off. Capturing them as structured records at the moment they occur is what makes the accessorial collectible rather than theoretical.

Prevention and capture run on the same infrastructure. The system watching ETA against appointment is the system that knows when the truck arrived.

What to measure

• Detention incidents per 100 loads, by facility. The diagnostic that identifies which consignees are systematically costing you — and that is a commercial conversation worth having with data.
• Average detention duration, and the share exceeding the free window.
• Rebooking rate. How often a mismatch was caught and fixed before arrival. This is the prevention metric, and most operations start at effectively zero.
• Driver notification rate. The share of waiting drivers who were told anything.
• Detention billed versus detention incurred. The capture gap.

That third metric is the one worth building toward. Everything else measures how much detention you had. Rebooking rate measures how much you prevented.

Frequently asked questions

How much detention is preventable? The share caused by appointment drift and stale scheduling is substantial and is addressable through earlier detection. Genuine capacity failures at the facility are not preventable by communication, but they are a minority of incidents.

Should we bill detention to customers? Consistently or not at all. Sporadic billing with no evident logic reads as arbitrary and damages the relationship more than the revenue justifies. Consistent, well-evidenced billing is normal commercial practice.

What does detention cost per hour? Industry reporting for 2025-2026 puts it at roughly $50 to $100 per hour after a two-hour free window, varying with equipment type and contract terms.

Does notifying a waiting driver actually help if they still wait? It does not reduce the hours, but it changes what those hours cost. A driver with accurate information can make hours-of-service and rest decisions deliberately instead of losing the day to uncertainty — and it materially affects whether that carrier takes your next load.

The bottom line

$15.1 billion a year, detention at 39.3% of stops, and the root cause is usually that two parties each knew their own schedule had moved and neither told the other.

Re-check every confirmed appointment against live ETA continuously, open the rebooking conversation the moment they diverge, tell waiting drivers where they stand, and capture arrival and departure timestamps so the residual is actually billable.

Debales deploys AI agents for freight quoting, order processing, ETA updates, and multi-channel customer communication — watching live ETAs against booked appointments and running the rebooking conversation before a driver arrives to wait. Book a demo.