Does buying AI mean replacing anyone on your pricing team?

The afternoon you tell the office you are looking at AI for quoting, one question travels the floor faster than any other. Not whether it works. Whether it is about them. By the second vendor meeting somebody has decided the answer, and you were not in the room.

That matters practically, not sentimentally. A pricing executive who thinks this is a headcount exercise feeds the trial the easy enquiries. Two cartons, a weekly lane, a customer who never argues. The awkward ones, the reason you are shopping, keep going into the old spreadsheet. Six weeks later the trial looks fine and tells you nothing.

Acas puts the general version plainly: employers should always consult workers and their representatives when making decisions, and it can help make better decisions and improve working relationships. Buying quoting software decides how somebody's day is spent. It qualifies.

Take one enquiry apart and the answer stops being an opinion.

An email lands at nine twelve. A customer's purchasing assistant wants a price for four pallets from Ho Chi Minh to Felixstowe, and the weights are in a photograph of a delivery note. Almost none of the next ninety minutes is pricing.

Eight steps. One is pricing. One more has pricing buried inside a search. The rest is finding, copying, waiting and chasing.

This is the split the question turns on. The International Labour Organization's 2023 working paper on generative AI and jobs reached the same conclusion for whole occupations: the overwhelming effect of the technology will be to augment occupations, rather than to automate them, with the greatest exposure in clerical work. The clerical part of that table is the hunting, and the hunting is not why you hired a pricing executive.

So the useful version of C3 is not “will anybody go”. It is “which of these rows does your product take, and which does it leave”. A vendor who takes six of the eight is selling a different thing from one who takes three. Both are legitimate purchases.

The honest answer is that she does the margin row of that table many more times a day, and almost none of the others.

• 01Her morning starts with a list rather than an inbox.
• 02Twelve enquiries came in overnight and eleven are drafted, each showing the rate it used, where it came from and what this customer paid in March.
• 03The first hour goes on those eleven, about a minute each, and on the twelfth, which has no rate anywhere and needs a call to an agent in Jebel Ali.

Four parts of the job grow, and they are the four that were being squeezed out before.

• The margin call. Not the arithmetic, the decision. This customer has three quotes open and a payment overdue, so this one goes out low and the credit conversation happens first.
• The exceptions. No rate anywhere, cargo that will not fit the equipment, a rule that changed last week. These used to be the interruptions. They are now the work.
• The agent and carrier relationships. A local charge comes back in an hour because somebody knows the person sending it. Software does not have that and should not claim it.
• Correcting the machine. When a draft is wrong, saying why in one sentence. Most vendors forget to mention this one.

There is a second effect worth knowing before you promise anything. In a customer support study published by the National Bureau of Economic Research, access to a generative AI assistant raised productivity by 14% on average, but 34% for novice and low-skilled staff, with minimal effect on the most experienced. Your newest coordinator gets quick fast. Your strongest pricing executive gains less on speed, because speed was never her advantage. What she gains is reach.

So sell it internally on reach, not on speed. Promise your strongest pricing executive that she will be faster and she will test it, find she is not, and write the whole thing off.

One worked example, from the only vendor we can quote without asking, which is us. FreighAI's answer to C3 is: “Your team approves. They do not compose, forward, file or chase. The reply goes out in their name.” It names what moves and what does not, and it survives being repeated on the floor. Hold it to the standard on this page, then make it prove itself with a reference your own size.

All four sound reasonable in the room and leave you with nothing a fortnight later.

Trap one. Higher-value work, with no noun after it

“It frees your team for higher-value work” is the commonest answer to C3 and it says nothing. Higher value than what, done by whom, in which hours. Push once: “Name three things my pricing executive will do next quarter that she has no time for today.” A vendor who has watched a real pricing desk answers straight away.

Trap two. Headcount arithmetic done in your room

The mirror image is worse. A vendor who tells you the product saves two heads has told you something they cannot know. They have not seen your enquiry mix, your rate sources, or how much of your quoting depends on agents. Ask what the number rests on. If it is a model rather than a named firm your size, it is a brochure figure.

Trap three. The answer you get is not the answer your team gets

Some vendors carry two versions. A reassuring one for the pricing desk and a headcount one for the person signing. Test it in a minute: ask them to give the same answer, in the same words, to the people who will use the thing. A vendor with one answer agrees on the spot.

Trap four. The same job, only faster

Some products take nothing off the desk. They make the hunting quicker: better search, a nicer rate screen, fewer clicks. That is a real product and can be worth buying. It is not what C3 asks about, and selling it as “you will stop hunting” gets you caught out in week three.

One tell runs across all four. Watch whether the answer uses the word at all. The checklist calls the warning sign “anything that avoids the word”. A vendor who says “no, and here is why”, or “yes, in one role”, is easier to buy from than one who talks for ninety seconds and lands nowhere.

Three lines, the same day, in the vendor's own words. You are writing something you will read out to your team, so write it the way you would say it.

Then put a date on it. Diarise a day ninety review against exactly those words, with the pricing desk in the room. Two measurements are enough: how many enquiries the same people handle, and how long a quote takes from arrival to sent.

Sometimes it is. The answer is still straight, just harder to give.

In freight it is rarely somebody leaving. It is a vacancy that stops being filled: an open coordinator seat carried eight months on overtime, and a decision not to recruit. Or a role moves department, and the person whose day was three-quarters chasing goes to customer service.

Not filling a seat is a different thing from emptying one, and your team can tell the difference. What they will not forgive is working it out three weeks in, when a job advertisement quietly disappears. If a role does change, say four things the day you announce the trial. Which role, what it becomes, when, and what happens to the person if the new shape is not what they want.

One last use for C3. Ask it about the system you already run. If somebody on the pricing desk left tomorrow and nobody took their seat, what would break first. If the answer is “the chasing”, you have found the thing worth buying without a vendor in the room.