From Rate Confirmation to Load in Seconds: How AI Load Entry Works

Every load starts with a rate confirmation: a PDF from the broker with the load number, pickup and delivery addresses, appointment windows, the rate, the commodity and a page of terms. Then someone types it all into the TMS, one field at a time.

That typing costs a few minutes per load and causes a steady trickle of mistakes: a transposed load number, a wrong appointment, a rate entered without the fuel surcharge. Those mistakes show up later, on an invoice the broker rejects or a truck that arrives on the wrong day. AI load entry removes most of the typing and leaves the dispatcher to check the result.

What the AI reads from a rate confirmation

Rate confirmations look different from every broker, but they all carry the same facts. A good AI load entry reads the document itself instead of relying on a fixed template, and pulls out:

• The broker or customer and the broker's load or order number.
• Every stop in order, pickups and deliveries, with facility names and addresses.
• Appointment dates and windows for each stop, in the facility's local time.
• The rate, including line haul and accessorials where they're listed.
• Miles, commodity, weight, equipment type and reference numbers (PO, BOL, pickup numbers).

In Ella, this is AI Load Con, included in the base system. You upload the PDF (or several at once), and Ella creates the load with its stops, the customer, the billing details and the document filed in the load's folder.

Why a dispatcher still checks it

AI is very good at reading documents and occasionally wrong about one. A smudged scan, a stop listed in an unusual place or a rate split across two lines can trip it up. That's why the result should be treated as a draft: the dispatcher opens the new load, checks the stops, dates and rate against the PDF, and fixes anything that's off.

Checking a filled-in load takes far less time than typing one, and the dispatcher's attention goes where it matters: whether the load is right, not where each field goes.

The awkward PDFs

Real rate confirmations aren't always clean. A practical system has to handle:

• Multi-stop loads with several pickups or deliveries, kept in the right order.
• Scanned or photographed confirmations, not just digital PDFs.
• Password-protected or damaged PDFs. Ella flags these and offers a fix instead of failing silently.
• Duplicates. If the same broker load number already exists, Ella stops and points you to the existing load instead of creating a second one.
• Several at once. Upload a batch at the start of a shift and watch each one turn into a load, with a clear message for any that need attention.

Straight from the inbox

Most rate confirmations arrive by email, so the fastest path skips the download entirely. With Ella's Gmail Suite, each person's work Gmail opens inside the TMS. When an email carries a rate confirmation for a load that isn't in the system yet, an Import load button appears. One click takes you to the loads page and starts the AI import, and the email thread stays linked to the new load.

Rolling it out to your dispatchers

1. Start with one or two dispatchers and your most common brokers.
2. Agree on the check: stops, appointment times, rate and broker load number, every time.
3. Watch for patterns. If one broker's format causes the same correction again and again, that's worth reporting so it can be handled.
4. Once the team trusts it, make uploading the rate confirmation the first step of booking a load, so the document is always filed with the load.

AI load entry is one of several AI features in Ella. It's also one of the easiest to measure: count the minutes per load before and after.

Frequently asked questions