Verified pickups · Mobile, kiosk and portals
Four parties, one pickup. A facility, a carrier, a driver, and a consignee each hold a different piece of the same event — and the person who loses the freight is the one who was never in the room. We built the room.
The situation
A truck arrives at a loading dock. A driver hands over a pickup number. Someone at the gate writes down a name and a plate, the trailer is loaded, and the truck drives off. It takes twenty minutes and nobody thinks about it again — until a load never arrives, and a carrier nobody hired turns out to have collected it.
That is double brokerage, and the reason it works is structural. The shipper's customer hires a broker, the broker assigns a carrier, that carrier quietly re-sells the load, and a different truck shows up at the dock. Every party has a record. None of them is the same record, and the only person physically present is the driver, who is often the last to know they are part of a scheme.
The facility could not verify what it could not see. Whether the driver held the licence they presented, whether the truck belonged to the carrier on the paperwork, whether that carrier was the one actually assigned — those were questions for an investigation afterwards, not a gate at nine in the morning. And the company whose freight it really was, the consignee, found out last of all.
The actual problem
A pickup is a claim made by whoever is standing at the gate. It becomes a fact only when someone else can check it.
Starting point
The client is a carrier verification company: it already knew more about the trucking industry's carriers than almost anyone, with warehoused FMCSA records, insurance history, ownership signals and a driver identity network. What it did not have was a presence at the one place that data would actually change an outcome — the loading dock, at the moment freight changes hands. Facilities were running the gate on paper and memory, and the verification data sat in another building.
How it started
The gate is a hostile environment for software: rain, gloves, bad signal, a driver who wants to leave, and an operator who will abandon any step that takes longer than the old way. So the work started at real docks, watching real check-ins, before a screen was drawn. The question was never "what should the app do" — it was "what does the operator have in their hands, and for how many seconds".
We timed check-ins at facilities that ran the gate differently — by appointment, by paperwork, by a guard who recognised drivers — and mapped what each one actually verified versus what it believed it verified.
Facility staff, carrier dispatchers, drivers, and the consignees who never see the gate. The most useful finding: the consignee wanted evidence, not notifications.
A licence scan happens once. The driver may already be walking away. That single constraint shaped the whole flow — nothing captured at the gate can be lost to a dropped connection, a closed app, or a wrong tap.
Working screens on real devices at real docks, early. If a step could not be completed one-handed in daylight glare, it was redesigned.
Architecture
The centre of the system is a single load release: one truck, one visit, one pickup, with every photo, scan, decision and timestamp attached to it. The facility writes to it at the gate. The carrier's roster and the driver's own identity profile are read against it. And the consignee — the company the freight belongs to — sees it the moment it exists, because the record carries their approval of the facility as a condition of its own creation.
What moved in
The process at a facility takes a driver from arrival to release in a sequence that adapts to that facility's rules — which steps are mandatory, which can be skipped, whether the truck or the driver is checked first — without becoming a different app.
A driver arrives, enters a phone number, pickup number and carrier on a tablet, and consents to identity verification. The visit exists before any staff member has touched it, and the carrier is already resolved against public records.
The barcode on the back of the licence, or the driver's own QR code from the driver app, resolves to a known driver and their current employer. Staff confirm the face against the profile photo. Unknown drivers go through a hosted government-ID check that reports exactly what failed — altered document, face mismatch, photo of a screen.
A photo of the truck reads the plate and its state automatically. Known equipment resolves to a fleet record with VIN and carrier. Unknown equipment requires a second photo of the cab door, from which the USDOT number is read and checked against the carrier the plate was last inspected under.
If the driver's carrier and the truck's carrier are different companies, check-in is blocked. It is an identity problem, not a paperwork problem, and the gate cannot wave it through.
Facility-defined document requirements, captured as galleries or guided step-by-step prompts, split between check-in and release. The load cannot be released until the server confirms every required document exists.
Gate assignment and completion notices sent by text from editable templates, each stamped with who sent it and when. The driver knows where to go without a radio call.
One dashboard for both sides. The facility sees its own releases; the consignee sees every load released on its behalf at any vendor's gate — same evidence, same photos, same verdicts.
Pickups by verification type, carriers by volume, check-in to release handling time per facility, team activity by operator, six-month trends. All derived from what the gate captured, not entered separately.
Connected systems
The gate is only as good as what it can check against. Most of the value comes from connecting the capture at the dock to records that already existed elsewhere.
What changed
The measures below were reported by the client from facilities running the system. Speed mattered because a slower gate would have been abandoned; the fraud figure mattered because it is the reason the system exists.
Time from driver arrival to completed check-in, halved against the paper process at the same facilities. Kiosk self-entry, automatic plate reading and known-driver recognition removed most of the typing.
Multiple attempted pickups by trucks or drivers that did not belong to the carrier on record were blocked before loading. Previously these were discovered as cargo claims, weeks later.
Shipper and consignee now look at the same record: verdicts, photos, documents and handling time for every release, including consignees with no facility of their own.
Three things at once. First, no two facilities run a gate the same way, so the workflow itself had to be configurable per facility — which steps, in which order, how strict — without forking the app. Second, the gate is offline-hostile: a licence scan is a one-shot capture, and nothing could be allowed to depend on a connection. Third, verification is only useful if it is decisive: a bare 'not verified' cannot be acted on, so every check had to say what failed and whether that failure was a missing step or a positive signal against the driver.
Native mobile for the yard, a public kiosk for the driver, and web portals for every party over one shared record. Identity and equipment resolved against existing carrier, driver and FMCSA data rather than re-entered. Consignee approval built into the record so visibility follows ownership automatically. Every policy decision — what to require, what to skip, whom to trust — pushed down to the facility and the freight owner, not hard-coded.
Scope
Working principle
Verify at the moment it can still change the outcome. Everything after that is an investigation.
Start with the operation
Every engagement starts the same way, whatever the industry: understanding how the work actually happens.
Show us how it works →