How a Regional Distributor Stopped Finding Delivery Problems at Billing
Use case Fuel delivery managementRegional fuel distributor, 42 trucks
How a regional distributor stopped finding delivery problems at billing.
Late loads, short drops and missing tickets used to surface days later, as held invoices and angry dealers. One sentence to Cortex built an assistant that catches them the same hour.
01 · The challenge
A morning spent chasing paper.
About one load in sixteen had something wrong: a late arrival, a partial drop or a missing delivery ticket. Nobody knew until accounting tried to invoice it. By then the driver had forgotten and the dealer was already annoyed.
Every morning, two dispatchers worked through yesterday’s loads across the dispatch system, driver tablets, emailed tickets and terminal portals. Closing one exception took nine manual steps.
Before Cortex
What the team saw first
- ExceptionsAbout 1 load in 16
- When foundAt billing, 3–4 days later
- Records4 systems, none matched automatically
- Cost~25 dispatcher hours a week
The difference
Same question. Different results.
The question“Which deliveries need attention right now?”
Claude / GPT / Public LLMs
Uploaded file
Answer
About 1,000 loads were delivered last week. 64 have notes in the log. Review them with dispatch.
Just an answer.
Cortex AI
Connected intelligence
Cortex AI- BOLs
- Delivery tickets
- Terminal status
- Weather
- Road closures
- Truck GPS
- Supplier invoices
- Tank levels
- Road traffic
- Allocation alerts
- 3 exceptions overnight
- Storm due in 48 hours
- Terminal on allocation
Answer
Three loads need action now: one late, one short drop and one missing ticket. Each task is assigned with the records attached. A storm is due Wednesday, so I’ve flagged 37 sites to load early.
A decision.
02 · A day with Cortex
Built in one sentence. Live the same week.
Cortex turned that sentence into a workflow. It watches every load and stays quiet until something is wrong. When it is, the right person gets a task with the records already attached.
The real test came in January. A winter storm hit while one terminal went on allocation. Cortex combined dealer tank levels, the storm forecast, road closures and terminal availability, and flagged the sites at risk a day and a half early.
- The assistant checks overnight loadsThree exceptions: one late, one short drop, one missing ticket
- Tasks assigned, records attachedDispatch owns two, accounting owns one
- All three closedInvoices release once the records match
- Storm forecast, terminal on allocation37 sites flagged as short on cover for the storm
- Loads pulled forward from the second terminal4 emergency loads, versus 19 in the last big storm
03 · The evidence
The assistant, and the week it mattered.
Every number traced back to its source record, so each fix had an owner and proof.
What Cortex found
The assistant, and the week it mattered
Delivery exception workflow
Watches dispatch records, BOLs and delivery tickets. Acts only when something is wrong.
- Monitor dispatchLoad, route and scheduled window
- Verify deliveryBOL matched to the delivery ticket, gallon for gallon
- Detect exceptionDelay, shortage or missing document
- Assign taskRouted to dispatch, records attachedAlert operationsException and next action
- Confirm resolutionCloses only when the records reconcile
The January test
Signals combined: dealer tank levels · storm forecast · road closures · terminal on allocation
Pull the 37 at-risk loads forward and lift them from the second terminal, before the allocation tightened.
- Dispatch now approves one or two steps per exception instead of doing nine.
- Invoices release on their own once the records match.
- Every exception leaves an audit trail with cause, owner and fix.
04 · The impact
What changed.
Exceptions are cheapest the hour they happen.
We used to hear about a short drop from the dealer’s angry phone call. Now we know before the driver is back at the yard.
A delivery problem fixed the same day is a phone call. Found at billing, it is a dispute, a held invoice and a dealer who starts shopping around.
