Two hours of copying before planning can start
At half past three every afternoon, the orders for tomorrow arrive. A building products customer sends a spreadsheet with forty lines. A packaging customer sends a PDF per load. A smaller customer writes an email: 'Can you collect 12 pallets from us tomorrow to the usual place in Wakefield, plus 4 to the new site'. Somebody in the office types all of it into the transport system, line by line, and planning cannot really start until they have finished.
Mistakes creep in: a wrong postcode, a missing reference, a pallet count transposed. They are usually found when the driver is at the wrong gate.
Why it has not been fixed already
The customers will not change how they send orders, and your transport system expects its own format. Retyping is what fills the gap.
- Every customer's format is different and they change them without warning.
- Some orders refer to 'the usual place' rather than an address.
- References the delivery site needs are buried in customer paperwork.
- Your transport system's import, if it has one, expects a strict template nobody sends.
- Late amendments arrive as replies to the original email.
What retyping costs
It is the office's time every afternoon, at exactly the point planners need the orders. Typing errors turn into failed deliveries, wasted journeys and customer complaints. Late amendments that are missed cause the same. And because it is tedious, it is hard to keep good people doing it. The person who knows each customer's quirks, that one always forgets the booking reference and another puts the delivery date in the collection column, becomes someone the office cannot manage without, and the afternoon stalls when they are off.
The order intake we build
- Orders arriving at your orders inbox, or uploaded to a customer web form, are picked up automatically.
- An extraction step, set up for each customer's format, reads spreadsheets, PDFs and emails and pulls out collection and delivery points, dates, times, references, pallet counts, weights and instructions. For free text emails, a language model does the reading, and every field it pulls out is checked.
- Phrases like 'the usual place' are resolved against that customer's known addresses. Postcodes are validated.
- Each order is checked against what you expect: known customer, valid dates, sensible weights, required references present. Anything odd is flagged.
- Clean orders are created as jobs in your transport system through its API or import. Flagged ones wait in a queue with the original document beside the extracted data.
- Amendment emails are matched to the original order, and the change is shown for approval.
| Customer format | Today | With order intake |
|---|---|---|
| Spreadsheet | Copied line by line | Read and loaded |
| PDF order | Retyped | Extracted and checked |
| Free text email | Interpreted and typed | Read, resolved, checked |
| Amendment | Easy to miss | Matched to the original |
The afternoon afterwards
Orders reach the transport system as they arrive rather than in an afternoon batch, and the office deals with the few that need a question answered. Planners start planning earlier. Wrong addresses and missing references are caught before the driver finds them. The orders inbox stops being a job in itself. The person who used to retype orders moves to checking the exceptions and talking to customers about the ones that are unclear, which is a better use of what they know. Over time, the queue shows which customers send the most problem orders, which gives you something specific to raise with them.
If your transport system cannot take jobs automatically, we tell you that at the start, and the intake produces a clean import file instead.
Checklist
- Someone spends hours a day retyping orders.
- Planning waits for orders to be entered.
- Deliveries fail because of typing errors.
- Customers send orders in many formats.
- Amendments get missed.