Blog

Should AI Recommend the Customer Rate When You Quote a Load?

By Connor Miller, Founder and CEO, FreightScout. Last updated 28 October 2026.

Short answer: yes, but only if you can see where the number came from. A rate recommendation is worth using when it's built off the lanes you've actually run, your cost on those lanes and the current buy, when you can open it and trace it to the loads behind it, and when a person approves it before it goes to the shipper. If a tool hands you a number and can't tell you why, don't send it. And on a lane you've never run, the right answer from any tool is "I don't know," not a guess.

Why do brokers roll their eyes at AI pricing?

Ask on a broker forum whether anybody uses AI to bid freight and recommend the customer rate, and the replies usually come back the same way. Not a chance. The rate's the one thing I'm not handing off.

That reaction makes sense. Everybody on a desk has been burned by a number somebody else came up with. A rate tool that said a lane was soft the week capacity dried up. A national average that had nothing to do with your shipper's dock, your carriers or the way that lane actually moves. You quote it, you win it, and then you spend the next two days covering it upside down.

So the objection isn't really "AI can't price freight." It's "I don't trust a number I can't check." That's the right instinct. It's the same reason you wouldn't let a brand-new rep send a quote on a big account without looking at it first.

What a rate recommendation is actually for

Pricing a load isn't one decision. It's a pile of small jobs, and most of them aren't judgment calls at all:

Reading the quote request out of the inbox and figuring out what's actually being asked
Pulling what you've paid on that lane, and when
Checking what it costs to buy a truck on it right now
Writing the reply back to the shipper
Logging the quote so the next one on that lane isn't starting from scratch

The judgment part is the margin you put on top, and whether this shipper, this week, is worth sharpening your pencil for. That part's yours. Everything else is busywork that eats the first twenty minutes of every quote, and the broker who answers while the freight's still open is the one who's in the conversation.

A good recommendation does the busywork and hands you a number with the work shown. You make the call.

What should a rate recommendation show you?

Here's the test. Before you trust any recommended rate, from software or from a rep, it should be able to answer these.

Question to ask	What a good answer looks like	Red flag
Where did the number come from?	Your own lane cost history and the current buy rate, or the quote tool you already pay for	A market average with no link to your book
Can I trace it?	You can open the figure and see the loads behind it	"Trust the model"
What if we've never run the lane?	The gap is marked and a person fills it	A confident number anyway
Who sends it?	You approve the reply before it goes out	It goes straight to the shipper
Where does the quote live after?	Logged in your TMS, so the next quote on that lane sees it	Stuck in someone's inbox or a side app
Spot or contract?	Both priced off the same book	Two tools that don't agree with each other

If a tool can't pass that table, it's not a recommendation. It's a guess with a nice font.

Your book is the best rate data you've got

Here's something I learned running a floor. The most accurate pricing information in the building wasn't a subscription. It was our own history. What we paid on that lane last month. Which carriers ran it and what they took. What the shipper accepted and what they pushed back on.

The problem was nobody could get to it fast. It lived in the TMS, spread across a hundred loads, and pulling it together for a single quote took longer than the shipper wanted to wait. So reps guessed, or they quoted off memory, or they quoted off whatever the last guy on the desk said.

That's where a rate recommendation earns its keep. Not by inventing a better number than your book. By putting your book in front of you while the quote request is still fresh.

Spot and contract should come off the same book

Spot is this load. Contract is the lane over time. They fail the same way when the price is a guess.

If your spot quotes come from one place and your bid file comes from somewhere else, you'll end up with a shipper who sees two different numbers for the same lane from the same broker. That's a bad conversation to have.

Price both off the same history: the lanes you've actually run, your cost and the current buy. Then when the RFP comes around, you're not starting a spreadsheet from zero. You're looking at the same record you've been quoting from all year. More on the bid side is in AI RFP response for freight brokers, and the basics of each are in the glossary under spot quote and contract rate.

What stays with your people?

Some things shouldn't be recommended to you at all. They should just be yours.

The price on a strategic account. If your name's on the relationship, your name's on the number.
Any number over max pay. It goes to a person, every time.
Lanes you've never run. No history means no recommendation. A rep fills that one.
The call on whether to chase the freight. Some loads aren't worth winning. That's a judgment call, not a lookup.

The way I'd put it to a younger broker: let the system bring you the history. You make the call when the account needs your name on it.

How to start without betting the book on it

You don't have to flip a switch and let software price everything tomorrow. Start the way you'd start a new rep.

Have it draft, not send. Every quote comes to a person to approve before it goes to the shipper.
Check the work on lanes you know cold. Pick the lanes where your team already knows the right number. If the drafted rate matches what your best rep would've said, and you can see why, that's how trust gets built.
Watch what it does on thin lanes. The lane you've barely run is the real test. If it marks the gap instead of filling it with a made-up number, that's a tool you can work with.
Keep the record in the TMS. Every quote, won or lost, should land where the rest of the book lives. That's what makes the next recommendation better informed.

That's it. No big rollout. Just a faster first draft that a person still signs off on.

What does FreightScout do here?

FreightScout is built by a broker, not a software company. It's an AI execution layer that works the load on the TMS you already run. It's not a TMS and not a load board.

Here's what it does on quoting, and nothing more than what's on our site:

When a spot request hits the inbox, it reads it and prices the load off your own lane cost history and the current buy, or off the quote tool you already use.
The figure is one you can open and trace to the loads behind it.
It drafts the reply in your voice with the price you're willing to send. You approve it before anything commits. A number over max pay goes to a person.
It doesn't invent a rate. A lane you've never run is marked, not filled with a made-up number, and a person fills the gap.
The quote's logged in the TMS your team already logs into, so the next quote on that lane can see this one. When the freight's won, Build, Cover, Track and Bill run on that same TMS.
Spot and contract quotes come off the same book.

Quoting is supervised by default, so nothing goes to a shipper without a person until you switch it. You can switch to autopilot per team and per action, and every action is logged and reversible. The price you'll defend on a strategic account stays with you.

It works with McLeod, Aljex and Tai, and it connects by API where the TMS offers one, and by reading it in place where it doesn't, set up during implementation. The TMS stays the system of record.

We don't publish a win rate or a margin number for quoting. The full rundown is on the freight quoting page, the history side is in AI freight broker quoting from lane history, and the inbox side is in AI spot quotes for freight brokers. If you're sizing up any AI tool, how to tell if freight AI does the work or only suggests it is a good place to start.

Frequently asked questions

Can AI recommend the customer rate on a freight quote?

Yes, and it's useful when the number comes from your own lane history and current buy rates and you can trace it to the loads behind it. It should draft the rate for a person to approve, not send it on its own.

Should I trust an AI rate recommendation?

Only one you can check. Ask where the number came from, whether you can open it and see the loads behind it, and what it does on a lane you've never run. If it can't answer those, don't send it to a shipper.

Where should an AI quoting tool get its rates?

From your book first: what you've paid on the lane and the current buy. If your team already uses a quote tool, it should plug into that. A national average with no link to your history isn't a recommendation for your shipper.

What happens on a lane I've never run?

The right answer is to mark the gap and let a person price it. FreightScout doesn't invent a rate. A lane with no history is flagged, not filled with a made-up number.

Does the AI send the quote to the shipper?

Not unless you switch it. FreightScout's quoting is supervised by default, so it drafts the reply and you approve it before anything commits. A number over max pay goes to a person.

Can the same tool price spot quotes and RFP bids?

It should, so the shipper doesn't see two different numbers for the same lane. FreightScout prices spot and contract quotes off the same book, and the quote is logged in your TMS.

Does FreightScout replace my TMS or rate tool?

No. FreightScout isn't a TMS and isn't a load board. It uses the quote tool you already have or your lane history, and it works on the TMS you already run, including McLeod, Aljex and Tai. The TMS stays the system of record.

All posts: the blog index.