Rate con to ready-to-bill: where a freight brokerage’s hours actually go

A 30-person freight brokerage can spend around 240 hours a month on the paperwork that follows a booked load. That is roughly one and a half full-time people. It’s my estimate, not a survey result, so below I show how I got there and how to check it against your own desk.

The work I mean isn’t sales or dispatch. It’s the quieter loop that runs after a carrier is booked: getting the signed rate confirmation back, reading the proof of delivery, matching the carrier’s invoice and deciding what to pay.

Nobody brags about this work, and most automation pitches skip it. It’s also where small errors quietly turn into lost margin.

The paperwork loop behind one load

Every load runs through the same six steps, and each one has its own way of going wrong.

1. The rate con goes out. A rep books the carrier and sends the rate confirmation from the TMS, with the lane, dates, rate and the rules for extras.
2. The signed copy comes back. Someone chases it, then checks that the signature, the rate and the carrier’s MC number match what was booked. A mismatched MC is easy to miss, and it is one of the signs of carrier identity fraud.
3. The POD arrives. Often it’s a phone photo of a creased BOL, sometimes rotated, sometimes with handwriting in the corner. Someone has to read it for the receiver’s signature, the piece count and notes like “2 pallets crushed”.
4. The carrier invoice arrives. It gets matched line by line against the rate con: line haul, fuel and anything extra.
5. Extras get approved or disputed. Lumper, detention, layover and TONU charges should only be paid when they were approved and documented. Each one needs a decision, and often an email.
6. The load moves to ready-to-bill. Only now can the customer be invoiced, so every delay above pushes cash further out.

Where the time goes

For a brokerage moving about 1,200 loads a month, my rough estimate is 12 minutes of paperwork per load, or about 240 hours a month.

These are estimates, not measurements. Teams with clean carrier portals will come in lower; teams working from shared inboxes and phone photos will come in higher. At about 160 working hours per person a month, 240 hours is one and a half people.

The costs that never show up as hours

The hours are the visible cost. The bigger ones hide in the gaps between steps.

• Extras that slip through. A $290 lumper charge with no approval and no receipt looks like any other line on a busy afternoon. Paid a few times a week, it becomes real margin.
• Slower billing. A load can’t be invoiced until the POD and the carrier invoice are clean. Every day it sits in someone’s inbox adds a day to getting paid.
• Fraud exposure. Carrier identity theft and double brokering often show up as small inconsistencies, like an MC number that doesn’t match the carrier on file. They’re easy to catch when every document gets checked, and easy to miss when nobody has time.
• Peak-season strain. When volume jumps, paperwork backs up first, because it’s the work nobody sees.

What I’d automate first

I’d start with invoice matching, because it’s where money leaks and the rules are already clear.

1. Invoice vs rate con matching. Read every carrier invoice, compare it with the agreed rate plus approved extras, and flag any difference with the reason and a drafted dispute email. Clean invoices go straight through.
2. POD reading. Pull out the receiver’s signature, the piece count and any exception notes, including handwriting on a phone photo, and flag anything that should stop billing.
3. Carrier identity checks. Compare the MC number on every signed rate con and invoice with the carrier on file, and hold anything that doesn’t match.
4. Load tenders into draft loads. Turn a shipper’s emailed tender into a draft load in the TMS, so a rep starts from a filled record instead of a blank screen.

In all four, AI reads the document but plain rules make the calls: does the total match, is the signature there, is this the MC on file. That keeps every flag explainable, and it means the system never invents a number.

What I wouldn’t automate

I wouldn’t let software pay, dispute or release anything on its own.

• Pay or dispute decisions. The system should put the evidence in front of a person, with the dispute already drafted, and let them click. A wrong automatic dispute can damage a carrier relationship that took months to build.
• Carrier relationships. Negotiating, calming a driver down, or making an exception for a carrier who always delivers is human work.
• Anything without an audit trail. If a flag can’t show which document and which line it came from, nobody should act on it.

Check your own numbers in one week

You can test my estimates against your own desk in five working days.

1. Count the loads you deliver in a normal week.
2. Ask whoever handles paperwork to jot down the minutes spent on rate cons, PODs and invoices each day. Rough is fine.
3. Count how many carrier invoices didn’t match the rate con, and by how much in total.
4. Note the average number of days between delivery and the customer invoice.
5. Divide the paperwork minutes by the loads. My rule of thumb: above about 8 minutes a load, there’s an easy win.

About me

I’m Dhruv Rana, founder of OnFrontBase. I build AI systems that read documents and messages inside the tools a business already uses, with a person approving anything that matters.

I’m a new, one-person agency, so I won’t claim results I don’t have. What I’m finishing this week is a demo of this exact loop on sample paperwork: signed rate cons, a phone-photo POD with a damage note, a carrier invoice with an unapproved lumper charge, and a load tender.

If you run ops or billing at a brokerage and want to see it, message me on LinkedIn and I’ll send you the 2-minute walkthrough. And if your numbers look nothing like mine, I’d like to hear that too.