Shipment tracking software: when should an overdue-delivery alert fire?
Key Takeaways
An overdue-delivery alert should fire when the promised delivery window closes and there is no verified delivery record, such as a delivery scan or a proof of delivery. A second, earlier alert should fire when a verified ETA shows the load will miss the window. An AI logistics coordinator should act on both alerts, not only display them.
- A delivery window is the date and time range the customer or receiver expects the freight.
- An overdue load is one whose window has passed with no verified delivery. A late-risk load is one whose verified ETA falls after the window.
- The two alerts serve different purposes. The risk alert gives time to warn people. The overdue alert starts a search for the freight.
- Alerts should be tuned by mode and lane, because LTL, truckload, drayage, and parcel report delivery differently.
- An alert is not a resolution. Each alert should lead to a named next step and owner.
- Wilson, by Cartage, is an AI logistics coordinator that tracks each load to delivery, follows up with the carrier when delivery is not confirmed, and notifies the people who need to know.
What is the difference between a late-risk alert and an overdue alert?
A late-risk alert fires before the delivery window closes, when a verified ETA shows the load will arrive late. An overdue alert fires after the window closes, when no verified delivery record exists. An AI logistics coordinator should treat them as two separate events with different actions.
Many tracking tools send one kind of alert: "late." That single label mixes two very different situations. In one, the truck is moving and will arrive three hours after the window. In the other, nobody knows where the freight is. The first needs a message to the receiver. The second needs a search.
The table below separates the two alerts.
| Feature | Late-risk alert | Overdue alert |
|---|---|---|
| When it fires | Before the window closes | After the window closes |
| What triggers it | A verified ETA after the window | No verified delivery record |
| What is known | Where the load is and when it should arrive | Delivery is not confirmed |
| First action | Tell the receiver and the internal team | Contact the carrier to confirm status |
| Main risk | Missed appointment, rescheduling | Lost or misdelivered freight, missing POD |
| Who owns it | Logistics team, customer service | Logistics team, then carrier |
Why "verified" matters
An ETA, or estimated time of arrival, can come from a calculation or from a fresh carrier update. A verified ETA comes from a fresh fact: a carrier message, a GPS or ELD location, or a terminal scan. An ELD, or electronic logging device, records a truck's location and the driver's hours. An alert based on an unverified ETA can fire too early or too late. An AI logistics employee should base the late-risk alert on the latest verified information and show its source.
Why two alerts beat one
Two alerts give the team two chances to act. The late-risk alert gives time to move an appointment, warn a plant, or tell a customer before the window closes. The overdue alert catches the loads that slipped through without warning. An AI logistics coordinator that runs both turns most late deliveries into planned changes rather than surprises. A single alert at the end of the window leaves only damage control.
Why delivery needs proof
Delivery is confirmed by evidence, not by the passing of time. The strongest evidence is a proof of delivery, or POD, which is the signed record that the receiver took the freight. A carrier's delivery scan is useful but weaker. A load that passed its window with a delivery scan but no POD may be delivered, but the record is not complete. The overdue alert should consider which evidence exists.
Early arrivals deserve attention too
An early arrival can cause problems as well. A receiver with a strict appointment may turn away a truck that arrives hours early, or make it wait and bill detention. Detention is a fee a carrier charges when its truck waits too long at a dock. Some teams add an early-arrival alert for strict appointments. An AI logistics coordinator can tell the carrier to hold or ask the receiver whether an earlier slot is open.
The cost of a single "late" label
When every problem is labeled "late," teams learn to ignore the label. A coordinator who sees 40 "late" flags each morning cannot tell which ones need a phone call. Splitting the alert into late-risk and overdue, and adding the evidence behind each one, lets the team act on the few that matter.
When exactly should each alert fire?
The late-risk alert should fire as soon as a verified ETA falls after the window by more than a small tolerance. The overdue alert should fire when the window has closed plus a grace period with no verified delivery. An AI logistics coordinator should apply these rules by mode, because each mode confirms delivery differently.
Example timing by mode
The timings below are examples for discussion, not industry standards. Each team should set its own.
| Mode | Late-risk alert fires when | Overdue alert fires when |
|---|---|---|
| Full truckload | Verified ETA is more than 30 minutes after the appointment | Appointment plus 1 hour passes with no delivery confirmation |
| LTL | Carrier's delivery date moves past the promised date | End of the promised delivery day passes with no delivery scan |
| Drayage | Delivery appointment at risk from terminal delay | Appointment plus 2 hours with no delivery confirmation |
| Parcel | Carrier's expected date moves past the promised date | End of the promised day with no delivery scan |
| Cross-border | Border handoff delay pushes ETA past the window | End of the window with no delivery confirmation |
Full truckload, or FTL, means one truck carries one shipper's freight. Less-than-truckload, or LTL, means several shippers' freight shares a trailer. Drayage is the short truck move of a container to or from a port or rail yard.
Where the delivery window comes from
The delivery window should come from the order or the appointment, not from the carrier's transit estimate alone. A sales order may carry a promised delivery date. A receiver may assign an appointment. An LTL carrier may quote a transit time that implies a date. An AI logistics coordinator should record which source set the window, because the alert is only as good as the window behind it. If the order says Friday and the carrier quote says Monday, the conflict should be raised before the load ships.
Tolerance and grace periods
A tolerance stops alerts for tiny delays. A truck 10 minutes late to a two-hour appointment window is not a problem. A grace period gives the carrier time to report a delivery that already happened. Many carriers update delivery status an hour or more after the driver leaves. An AI logistics employee should apply both, so alerts point to real problems.
Windows versus appointments
Some deliveries have a strict appointment, such as 8:00 a.m. at a retailer's distribution center. Others have a wider window, such as "any time Tuesday." The alert rules should use the right one. A strict appointment needs a tight tolerance. A day window can use the end of the day.
Multi-stop loads
A load with several stops needs a window for each stop. A truck that delivers the first stop on time can still miss the third. The AI logistics coordinator should apply the alert rules to each stop and show which stop is at risk. A single window for the whole load hides the stop that actually needs attention.
Customer-specific windows
Some customers have stricter rules than others. A retailer may charge fees for deliveries outside its appointment. A construction site may need materials before a crew arrives. The AI logistics coordinator should apply the customer's window, not a generic one, when the customer is known.
Hypothetical example: a risk alert that saved an appointment
This example is hypothetical and does not describe a real customer. A snack manufacturer has a truckload due at a grocery distribution center at 6:00 a.m. Thursday. On Wednesday at 4 p.m., the carrier reports the truck is delayed by weather and now expects 9:30 a.m.
The verified ETA is three and a half hours after the appointment. The late-risk alert fires. The AI logistics coordinator tells the manufacturer's customer service team and asks the distribution center for a later slot. The center offers 10:00 a.m. The coordinator confirms the new time with the carrier. The truck arrives at 9:50 a.m. and delivers on time for the new appointment.
What should happen after an overdue alert fires?
After an overdue alert fires, the software should check every source for delivery evidence, contact the carrier, and resolve the status as delivered, still in transit, or a problem. An AI logistics coordinator should not close the alert until one of those outcomes is verified.
A practical overdue workflow
Check the sources. Look for a delivery scan, a POD, a GPS or ELD position, or a terminal event.
Contact the carrier. Ask whether the load delivered, and if not, where it is and when it will arrive.
Contact the receiver if needed. For some loads, the receiver can confirm delivery faster than the carrier.
Classify the outcome. Delivered with evidence, in transit with a verified new ETA, or a problem such as a refused or lost load.
Act on the outcome. Request the POD, update the ETA, or start an exception process.
Notify. Tell the internal team and, where configured, the customer.
Record. Save each step, source, and time.
When the carrier does not answer
Some carriers do not answer the first overdue follow-up. The rule should set a second step: a phone call, a message to a different contact, or a check with the receiver. If the status is still unknown after a set time, the alert should reach a manager. An AI logistics employee should record each attempt, so the manager sees what was tried. Silence after an overdue delivery is a serious signal and should not wait until the next day.
Requesting the POD
A POD request should be short and specific. For example: "Please send the signed POD for PRO 4418823, delivered to ABC Hardware on Tuesday." Carriers often keep PODs in an online portal or send them by email. An AI logistics coordinator can request the POD, check that it is signed and matches the load, and save it with the shipment. A POD is also needed later to release payment and resolve disputes.
Delivered but not recorded
Many overdue alerts turn out to be delivered loads with late paperwork. The carrier delivered on time but has not updated its system or sent the POD. The fix is to request the POD and record the delivery time. An AI logistics employee should still close the alert only when evidence arrives, because "the driver said it delivered" is not a POD.
Still in transit
Some overdue loads are still moving. The carrier may have missed a transfer or had a breakdown without reporting it. The AI logistics coordinator should get a verified new ETA and tell the receiver. The alert then becomes a late-risk event with a new time.
Partial deliveries
Sometimes only part of the shipment arrives. The receiver signs for eight pallets of ten. The POD shows a shortage. The alert should not close as "delivered" when the evidence shows a short delivery. An AI logistics employee should flag the shortage, ask the carrier where the missing pallets are, and tell the internal team. A partial delivery is an exception, even though a POD exists.
A real problem
Some overdue loads have real problems. The receiver refused the freight. The carrier delivered to the wrong address. The freight was damaged or lost. These cases need an exception process with a named owner. When Wilson detects a delay, missed pickup, or cancellation, it contacts the carrier, notifies the internal team, and notifies the customer where configured. Refused, misdelivered, or damaged freight may also need a claim, which a person should own.
Hypothetical example: the overdue load that was delivered
This example is hypothetical and does not describe a real customer. A distributor's LTL shipment to a hardware store was due Tuesday. At 8 p.m. Tuesday, the overdue alert fires because the carrier's tracking shows "out for delivery" and nothing more.
The AI logistics coordinator emails the carrier and checks the carrier's website by PRO number. A PRO number is the tracking number an LTL carrier assigns to a shipment. Wednesday morning, the website shows "delivered Tuesday 3:12 p.m." The coordinator requests the POD, which arrives that afternoon with the store's signature. The alert closes with the outcome "delivered, late paperwork," and the record shows the original delivery time.
Who should receive each alert, and how should alerts be tuned?
Each alert should go to the people who can act on it, with only the information they need. An AI logistics coordinator should route late-risk alerts to people who manage the receiver and overdue alerts to people who manage the carrier.
Tell people what changed, not only that something changed
An alert should state the change in plain words: the old time, the new time, and the reason if known. A receiver who reads "ETA now 9:30 a.m., was 6:00 a.m., weather delay, confirmed by carrier at 4:10 p.m." can decide quickly. An AI logistics coordinator should write every alert this way.
Routing by alert type
| Alert | Internal recipient | External recipient | Channel example |
|---|---|---|---|
| Late-risk, small delay | Logistics team | Receiver, if appointment is affected | Slack, Teams, or email |
| Late-risk, large delay | Logistics and customer service | Customer and receiver | Email or SMS where configured |
| Overdue, likely paperwork | Logistics team | None until resolved | Team channel |
| Overdue, unknown status | Logistics team and manager | Customer if delay is confirmed | Team channel and email |
| Overdue, real problem | Logistics manager and exception owner | Customer with a plan | Direct message and email |
Wilson sends internal updates through channels such as Slack, Microsoft Teams, or email, and can notify consignees by email or SMS where configured. A consignee is the party receiving the freight.
Tuning to reduce noise
Too many alerts teach people to ignore them. Review alerts after the first month. If most overdue alerts on one carrier are late paperwork, lengthen the grace period for that carrier or check its tracking source. If late-risk alerts fire for tiny delays, widen the tolerance. An AI logistics employee can show how often each alert fired and how each was resolved, which makes tuning easier.
Quiet hours and weekends
Overdue alerts that fire at 11 p.m. on a Saturday may reach nobody. The rule should say what happens outside working hours. Some teams hold non-urgent alerts until the next business morning. Others route urgent loads, such as production-critical inbound freight, to an on-call person. An AI logistics coordinator can keep following up with the carrier during quiet hours where configured, so the morning team starts with the facts.
Alerts and dashboards are different
A dashboard shows status when someone looks. An alert reaches people when something changes. Both are useful. The important point is that an alert should come with the first action already taken. A message that says "Load 3341 overdue; carrier contacted at 8:05 p.m.; waiting for reply" is more useful than "Load 3341 overdue."
Who owns alert rules
The logistics manager should own alert rules, with input from customer service, which hears from customers, and from operations, which plans receiving. The AI logistics coordinator applies the rules and reports how they perform. Changes should be recorded, so the team can tell whether a drop in alerts came from better service or from looser rules.
Common tuning mistakes
- One rule for every mode. LTL and parcel alerts fire too often; truckload alerts fire too late.
- Alerts from unverified ETAs. Calculated ETAs trigger false risk alerts.
- No grace period. Overdue alerts fire before carriers have time to report delivery.
- No owner. Alerts reach a shared inbox and nobody acts.
How should you evaluate shipment tracking software on overdue alerts?
Evaluate shipment tracking software by running one load that will miss its window and one load that delivered without paperwork, and watching what each alert does. An AI logistics coordinator should separate the two cases and resolve each with evidence.
Demo tests for overdue alerts
| Buyer requirement | Demo test | Evidence to see |
|---|---|---|
| Two alert types | Run a late-risk load and an overdue load | Different alerts, different actions |
| Verified ETA basis | Use a recalculated ETA with no new fact | No risk alert from the guess alone |
| Mode-specific timing | Run LTL and truckload loads | Different thresholds applied |
| Grace period | Deliver a load with a delayed scan | No false overdue alert |
| Evidence to close | Mark a load delivered verbally | Alert stays open until POD or scan |
| Routing | Trigger a large delay | Right people notified in the right channel |
| Record | Close an overdue alert | Sources, contacts, and outcome saved |
Demo questions about overdue alerts
- How does the AI logistics coordinator know the delivery window for each load?
- Which sources count as verified delivery evidence?
- Can thresholds vary by mode, lane, carrier, and customer?
- What happens when the carrier does not answer an overdue follow-up?
- How are alert counts and outcomes reported for tuning?
- Can the customer receive a late-risk notice before the window closes?
Test a carrier with slow updates
Run the overdue test with a carrier that updates its tracking slowly. Fast-updating carriers make every alert look accurate. A slow carrier shows whether the grace period and the follow-up steps work. The AI logistics coordinator should avoid a false alarm when the delivery happened and still catch a real problem when it did not.
How to measure overdue alert work
Count how many late-risk alerts fired before the window closed. Count how many overdue alerts were resolved with evidence within one business day. Track how many deliveries were missed with no prior alert. These measures show whether the AI logistics employee uses alerts to change outcomes.
Warning signs in an overdue alert demo
- Every delayed load shows the same "late" flag.
- Alerts close when the delivery time passes, with no evidence.
- There is no grace period for late carrier updates.
- Alerts list problems but show no action already taken.
- The vendor cannot show how many alerts were false alarms.
Scoring the overdue alert demo
Score each test as pass, partial, or fail. A pass means the right alert fired at the right time and the AI logistics coordinator resolved it with evidence. A partial means the alert fired but a person had to chase the carrier. A fail means the load passed its window without an alert, or the alert closed with no evidence.
Where does Wilson fit in overdue-delivery alerts?
Wilson, by Cartage, is an AI logistics coordinator that tracks each load from booking to delivery and follows up with the carrier when delivery is not confirmed. It replaces the TMS for freight execution and connects to the company's ERP through a method that fits the customer's environment.
Wilson tracks each load using the best source available, including carrier APIs, GPS or ELD data where it exists, the carrier's website by PRO number, and email or direct outreach, including calls where configured. This AI logistics employee asks the carrier for status and informs the people who need to know.
Delivery evidence and documents
Wilson collects shipping documents as part of each load, such as bills of lading and proof of delivery. A bill of lading, or BOL, is the shipping document that lists the freight and serves as a receipt between shipper and carrier. Wilson keeps searchable shipment records with change history, so a team can see when delivery was confirmed and by which source.
What Cartage asks about overdue alerts in discovery
Cartage asks how the team learns about late deliveries today and who needs to hear about them. Typical questions cover delivery windows by customer, which carriers report delivery quickly, and who owns refused or damaged freight. These answers shape the first alert rules an AI logistics coordinator applies.
What Wilson does not do for overdue alerts
Wilson does not promise a fresh location for every load at every moment. The freshness of updates depends on the source each carrier offers. Wilson is not a standalone visibility platform; it tracks loads as part of moving them. Specific alert thresholds and notification channels should be confirmed in a demo for the customer's setup. Exact ERP connection details depend on the customer's systems.
Where teams start with overdue alerts
Most teams start their AI logistics coordinator on a few lanes and expand from there. Timing depends on scope. Cartage helps identify the first workflow during discovery. Teams with strict retail appointments often start with those customers.
Teams that see alerts pile up can read why a delay alert is not a resolution plan. Teams chasing delivery paperwork can see when to chase a missing POD. Both topics shape what an AI logistics coordinator does after an alert. Teams testing customer messages can review how to test customer notifications for a late load.
FAQs
When should an overdue-delivery alert fire?
When the delivery window closes, plus a short grace period, with no verified delivery record. An AI logistics coordinator should also send an earlier late-risk alert when a verified ETA misses the window.
What counts as verified delivery?
A signed proof of delivery is the strongest evidence. A carrier delivery scan is useful but weaker. An AI logistics employee should keep the alert open until evidence arrives.
Why do many overdue alerts turn out to be delivered loads?
Carriers often update delivery status or send the POD hours after the driver leaves. An AI logistics coordinator should request the paperwork and record the real delivery time.
Should every overdue alert go to the customer?
No. Customers should hear when a delay is confirmed and affects their plans. An AI logistics employee should resolve paperwork gaps internally first.
Sources
- Cartage, Wilson product scope, October 2026: https://cartage.ai
Ask Cartage to show how an AI logistics coordinator separates a late-risk alert from an overdue alert, follows up with the carrier, and closes the alert only when delivery evidence arrives: talk to the Cartage team.