By Luis Lopez, AI transportation consultant, CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast
Detention is one of the oldest arguments in trucking. The driver says the truck waited for hours. The facility says the truck showed up late. The shipper says nobody told them. Without reliable timestamps, the conversation becomes a negotiation over memory, and the side with the better paperwork usually wins.
Geofencing changes that conversation. When arrival and departure times are captured automatically by location data, dwell time stops being an opinion and becomes a record. Add AI on top, and that record becomes something more useful: a way to predict delays, bill accurately, and fix the facilities that cause the most waiting.
What geofencing actually does
A geofence is a virtual boundary drawn around a location, such as a shipper’s dock, a warehouse yard, or a port terminal gate. When a truck’s GPS position, usually from an ELD, telematics unit, or driver app, crosses into that boundary, the system logs an arrival. When it crosses out, it logs a departure. The time between the two is dwell time.
That sounds simple, and the core mechanic is. The difficulty is in the details:
- Boundary placement. A fence drawn around a whole industrial park will count time spent in a neighboring lot. A fence drawn too tightly around a dock door may miss a truck waiting at the gate.
- GPS drift. Signals bounce near large buildings, so a parked truck can appear to exit and re-enter a fence several times.
- Staging areas. Many facilities make trucks wait outside the property, on a street or in an overflow lot. If the fence does not cover that area, the waiting is invisible.
- Appointment context. Dwell time alone does not tell you whether the truck was early, on time, or late for its appointment, which is usually what determines whether detention applies.
Where AI adds value
Raw geofence events are noisy. AI is useful for turning them into clean, defensible data and then into decisions.
Cleaning the events
Models can recognize GPS jitter and merge a burst of exit-enter events into one continuous stay. They can learn from historical tracks where trucks actually park and wait at a given facility, then suggest better fence shapes, including the off-site staging lots that matter most.
Matching events to loads
A truck may visit three facilities in a day. The system has to tie each geofence visit to the right load and stop. AI-assisted matching uses the planned route, appointment times, and stop sequence to attach each arrival and departure to the right shipment, and flags visits that do not fit any planned stop.
Applying the detention rules
Detention terms vary by contract: free time, when the clock starts, whether early arrivals count, billing increments. A rules engine reads the actual terms for that customer or carrier and computes billable time. AI helps by reading the terms out of rate confirmations and contracts so the rules are not typed by hand, with a person confirming the extracted terms before they are used.
Predicting long dwell before it happens
Once you have clean history, you can predict which appointments are likely to run long, based on facility, day of week, time of day, and load type. That lets dispatch plan the next load realistically, warn the customer early, and avoid building a schedule that breaks by noon. The broader topic is covered in AI dock scheduling and yard management.
Turning dwell time into data you can use
As an AI transportation consultant, the first thing I look at is what a team does with dwell data after it is collected. Many fleets and brokers have geofence timestamps already. Few use them beyond settling a single dispute. Here is where the value actually sits:
- Billing accuracy. Detention backed by automatic timestamps is easier to bill and harder to dispute. Attach the arrival, departure, and appointment time to the invoice.
- Driver pay. If drivers are paid for detention, automatic records remove the need for them to argue for every hour.
- Facility scorecards. Rank the facilities in your network by average dwell and variability. A small number of locations usually cause a large share of waiting.
- Customer conversations. Shippers often do not know how long trucks wait at their own docks. A clear, factual dwell report is a better starting point than a complaint.
- Pricing. Lanes that consistently involve long dwell carry real cost. That cost belongs in the rate, not in a dispute after the fact.
- Capacity planning. Hours spent waiting reduce what a truck and driver can do under hours-of-service limits. Dwell data shows where your capacity is actually going.
For practical steps on the operational side, see how to reduce truck detention time.
Setting it up the right way
Start with your highest-volume facilities
Draw careful fences around the locations you visit most. Include gates, staging lots, and the access road if trucks queue there. Review a sample of real truck tracks to confirm the fence captures where trucks actually wait.
Define the clock precisely
Decide, and put in writing, when the clock starts and stops: at fence entry, at check-in, at the appointment time, or whichever is later. Ambiguity here is the root of most detention disputes.
Keep the raw data
Store the original GPS points and events, not just the computed dwell time. If a number is challenged, you need to be able to show how it was produced.
Watch the data quality
Telematics gaps happen: devices lose power, phones go offline, drivers turn apps off. The system should flag missing data instead of guessing. The guide on questions to ask about AI, ELD, and telematics data covers what to check.
Limits and cautions
- A geofence proves where a truck was, not why it waited. A late arrival, a missing paperwork issue, or a mechanical problem can all produce long dwell that is not the facility’s fault.
- Location tracking of drivers raises privacy expectations. Be clear with drivers about what is tracked, when, and how the data is used.
- Port and terminal time often involves different charges and parties. Container detention and demurrage are not the same as truck detention; the explainer on demurrage vs. detention breaks down the difference.
Geofencing will not end detention disputes on its own. What it does is replace memory with records, and AI makes those records clean enough to bill from, predict from, and improve from. That is a meaningful shift for anyone who has spent a morning arguing about what time a truck pulled in.
For more conversations on freight and AI, subscribe to the Freight Guru Podcast.
About the author: Luis Lopez is a Miami-based AI transportation consultant and logistics entrepreneur, the CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast.
