By Luis Lopez, AI transportation consultant, CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast
Every broker and shipper knows the feeling. A load picks up, the tracking link shows a single ping at the shipper, and then nothing. The customer asks where their freight is, and the only honest answer is that nobody knows until someone reaches the driver.
Loads that go dark create work, anxiety and risk. Carrier tracking compliance, meaning the share of loads that produce usable location updates from pickup to delivery, is one of the most practical places to apply AI in a brokerage or shipper operation. This article walks through why tracking fails, how to measure it, and how AI can help close the gaps without treating every missed ping as a crisis.
Why loads go untracked
Before trying to fix tracking, it helps to know why it breaks. The causes are usually mundane:
- The tracking app was never installed or the driver did not accept the request.
- The ELD or telematics integration is not connected, or it is connected to the wrong truck.
- The driver’s phone is off, out of battery, or location permissions were turned off.
- Dispatch never passed the tracking request from the office to the driver.
- Coverage gaps in rural areas delay updates until the device reconnects.
- The driver changed mid-load and the new driver was never set up.
Most of these are process problems, not bad intentions. That matters, because the response should usually be a reminder or a fix, not an accusation.
Measure tracking compliance per carrier
As an AI transportation consultant, the first thing I look at is whether a team actually measures tracking compliance, or just reacts to the loads that go dark. Measurement should be per carrier, and ideally per dispatcher or driver where the data allows.
Useful measures include:
- Share of loads where tracking connected before pickup.
- Share of loads with regular updates through the whole trip.
- Longest gap between updates on each load.
- How quickly tracking was restored after a gap.
Over time, this produces a clear picture. Some carriers track nearly every load. Others connect only when pushed. That history should feed carrier selection and conversations, the same way on-time performance does.
Automated reminders that fit the load
A large share of tracking failures can be fixed with a timely nudge. AI-assisted workflows can send reminders at the right moments: when the load is booked, a few hours before pickup, at pickup if no ping has arrived, and whenever updates stop for longer than expected.
Good reminders are short, specific and sent through the channel the carrier actually uses, whether that is text, email or a voice call. Voice agents are increasingly used for this kind of routine contact; this piece on AI voice agents for check calls and carrier sales covers the trade-offs. Escalate to a person only when automated reminders do not work.
Inferring position from check-call text
Not every update arrives as a GPS ping. Drivers and dispatchers send messages like “just left the shipper, heading north,” “fueling near the state line,” or “two hours out.” Language models can read these messages, extract a likely location and status, and attach them to the load record with a clear label that the position is reported, not measured.
This keeps the load timeline useful even when electronic tracking fails. It also saves staff from copying notes by hand. The key is to keep reported and measured locations separate so nobody mistakes a driver’s estimate for a verified position.
ETA confidence, not just ETA
An ETA built on fresh GPS data is very different from one built on a check call from six hours ago. Predictive ETA tools should show confidence alongside the estimate: high when tracking is live and the route is predictable, lower when data is stale or the driver is approaching hours-of-service limits.
Showing confidence helps customer service teams set expectations honestly and focus attention on loads where the risk is real. For more on how these estimates are built, see predictive ETA and freight visibility explained.
When no tracking is a fraud signal
Most dark loads are innocent. Some are not. In freight fraud schemes such as double brokering or cargo theft, refusing tracking, disabling it right after pickup, or providing location data that does not match the route can be warning signs.
AI can help by combining signals rather than treating any single one as proof. A carrier with a new authority, recently changed contact details, a refusal to install tracking and a location that suddenly jumps off-route deserves immediate attention. A long-standing carrier with a dead phone battery usually does not. This overview of AI freight fraud detection and carrier vetting goes deeper on combining signals.
Privacy and driver consent
Tracking a load means tracking a person. Drivers should know what is being tracked, when tracking starts and stops, and who sees the data. Good practice includes:
- Getting clear consent through the app or the carrier before tracking begins.
- Tracking only for the duration of the load, not before or after.
- Sharing location only with parties who need it for that shipment.
- Keeping location history only as long as there is a business reason.
Respecting drivers is not just ethical; it also improves compliance. Drivers are more willing to share location when they trust it will not be misused.
Put tracking requirements in the rate confirmation
If tracking matters, write it down. The rate confirmation is the natural place to state that the load requires tracking, which method is accepted, when it must be active, and how updates should be provided if electronic tracking fails. Clear terms remove arguments later and give dispatchers something concrete to pass on to drivers. For a refresher on what this document should cover, read what a rate confirmation is in trucking.
A practical rollout
- Start measuring tracking compliance per carrier this month.
- Automate reminders at booking, pre-pickup and pickup.
- Capture position from check-call messages with a clear “reported” label.
- Add ETA confidence to customer-facing updates.
- Combine tracking behavior with other risk signals for fraud review.
- Put tracking terms in every rate confirmation.
None of this needs to be perfect on day one. The goal is fewer loads that go dark, faster recovery when they do, and better decisions about which carriers earn more freight.
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.