Implementation guide

How QR scanning and dispatch verification work

Scanning reads what each carton is; cameras at the dock see what is actually loaded. Together they check every item against the dispatch order and stop a wrong, short or excess shipment before the truck leaves. This guide walks through the flow from the first scan to the mismatch alert.

At a glance

How does dispatch verification with QR scanning work?

The dispatch order comes from the ERP or WMS. Operators pick with a scanning app that reads the barcode, QR code and printed label of each carton. At the loading point, an AI scanner camera identifies every item again as it is loaded and checks its SKU and quantity against the order, while a top camera follows the goods onto the truck. A match lets loading go on; a mismatch raises an alert and stops loading until it is fixed. Every dispatch is saved with a photo and timestamp. Vikbot, an AI warehouse automation company in Chennai, India, builds this flow from ScanBot, Dispatch Tracker and DockBot.

Method
Barcode, QR and label scanning at picking, plus AI cameras at the loading point
Checked against
The pick list, invoice or dispatch order from the ERP or WMS, or a spreadsheet
Catches
Wrong, short and excess items, before the truck leaves
Output
Verified dispatch, exceptions and photo proof per dispatch
Company
Vikbot, a brand of Clofus Innovations Pvt. Ltd., Chennai, India

Why scanning and cameras work together

A scan proves what a label says, at the moment someone scans it. It does not prove that the carton then went onto the right truck, or that nothing else went with it. A camera at the loading point closes that gap: it identifies each item as it is actually loaded and follows it onto the truck. Vikbot uses each where it is strongest:

  • Scanning at picking: ScanBot, a phone app, reads the barcode or QR code and the full printed label, including batch, MFG date and expiry, so the right item and batch are picked.
  • Camera checks at loading: Dispatch Tracker camera units at the loading or dispatch point, or DockBot over a short conveyor at the dock door, check every item against the dispatch order as it is loaded.

For where scanning fits at receiving, putaway and packing as well, see QR code and barcode scanning solutions for warehouses.

The verification flow, step by step

  1. The order arrives. The pick list, invoice or dispatch order comes from the ERP or WMS. DockBot can also run standalone from a spreadsheet. Where FleetBot runs on the dock CCTV, it reads the truck's number plate and fetches the pick list for that vehicle.
  2. Pick with a scan. The operator scans the QR code or label on each carton with ScanBot. The app reads the label, guides FIFO/FEFO picking by batch and expiry, and validates the item against the invoice, so a wrong item or an older batch is caught at the shelf.
  3. Check at packing, where there is a packing bench. PackBot uses a top camera on the box contents and a label camera on the box, and checks the items against the order before the box is closed.
  4. Identify each item at the loading point. As goods reach the loading station, the AI scanner camera identifies each item from its label, packaging and appearance and checks its SKU against the dispatch order. Dispatch Tracker also checks quantity and destination against the waybill.
  5. Follow the goods onto the truck. A top camera above the loading area tracks how goods move, so only items that passed the check are loaded.
  6. Match: loading goes on. A matched item is approved for loading. At DockBot a green signal allows it.
  7. Mismatch: alert and stop. An item that does not match raises an alert at once. DockBot shows a red warning and blocks loading; Dispatch Tracker can stop the conveyor.
  8. Close the dispatch. The dispatch is photographed and timestamped, and the verified result and any exceptions are sent to the ERP or WMS and to Fusion Insight.

The AI runs on a Vikbot-supplied edge computer on the site's own network, so checks carry on without an internet connection. ScanBot works offline and syncs when it reconnects.

What counts as a mismatch

  • Wrong item: an item or SKU that is not on the dispatch order for this truck.
  • Excess: more of an item than the order asks for.
  • Short: less of an item than the order asks for, flagged before the truck leaves rather than found by the customer.
  • Wrong batch or expiry: caught earlier, at picking, where ScanBot reads the batch and expiry from the label.

Proof and records

Each dispatch is saved with a photo and timestamp, so a dispute over what left the dock is settled from the record instead of a recount. Verified dispatches and exceptions reach the ERP or WMS and Fusion Insight; Vikbot WMS changes stock only when a movement has been verified.

Known limits

  • The camera checks what it can see. Items are checked one by one or carton by carton at a controlled loading point; a sealed carton is checked by its label, not its contents. Contents are checked earlier, at packing.
  • It knows the items it was trained on. The models are trained on the customer's own items, so a new item or pack design should be added before it ships at volume.
  • The order must be right. Goods are checked against the pick list or dispatch order; an error in the order itself is not caught at the dock.
  • Conditions matter. Lighting, glare and camera position affect reading, so they are checked at installation and during a parallel run alongside the existing checks.

What a site needs

  • A loading or dispatch point where goods pass in a controlled way, such as a conveyor to the truck or a dock door, with a mounting position, lighting and power checked during the site survey.
  • A place on the local network for the edge computer, and phones for ScanBot where picking is scanned.
  • The dispatch orders: a link to the ERP or WMS, or a spreadsheet for a standalone DockBot. SAP is connected over SAP APIs (OData, BAPI/RFC), with the SAP-side work done together with the customer's SAP team.
  • Samples of the items for model training, and staff for the parallel run.

A rollout follows the same steps each time: site survey, installation, model training on the customer's items, a parallel run and go-live. The timeline is set after the site survey. For how the cameras count, see how camera-based inventory counting works; for other guides, see all implementation guides.

Plan a site survey

Tell us how you check dispatches today, and we will show how scanning and camera verification would work at your dock.