The Mid-Market’s Next 24 Months: The AI-and-Analytics Bifurcation That Decides Who Keeps the Freight

A mid-market shipper watches its transportation cost creep up and can’t say why, so somebody sells them a shiny piece of software to fix it. A year later, the cost is still climbing, and now there’s a dashboard nobody opens.

I’ve watched that same movie since 1999.

So let me say what the software people won’t. AI is not going to save the mid-market. What it does is harder to sell and a lot truer: it splits us in two.

The companies that put it in the hands of people who know freight pull ahead. The rest buy the tool, wait on savings that don’t come, and slide backward a few points at a time, never feeling it until it’s structural.

That’s the whole reason NT Logistics exists. We don’t sell you a login and wish you luck. NTelligence finds where your freight is losing money, and then Freight Execution, our people, goes and gets it back. Finding and fixing under one team. Most outfits do one or the other.

I’ve spent 28 years and over $2 billion in freight learning which companies land on which side of that line. Now I’m taking the chair at the TIA, so I get to say it to a bigger room. 

Five calls I’d put money on for the next two years. Every one of them widens the same gap: the shippers who find the money and go get it, and the ones stuck doing half the job.

The Split Gets Visible: Distribution Cost Stops Being One Number

Everyone keeps calling AI the great equalizer. I think it’s the opposite. The operators who put analytics to work, finding the leak and then going after it, are pulling away from the ones who only bought the software. The numbers say so.

Coupa’s 2026 spend report found that leaders spot supplier and price risk early 60% of the time. Laggards manage 26%, and they lose revenue to failed orders 2.4 times as often. The profitability gap between the two groups runs about 23%. Supply Chain Management Review said it straight: AI is widening the distance between leaders and laggards.

Two competitors your size. Same trucks, same lanes, same customers. In two years, one of them knows its cost-to-serve cold, the true cost of moving each customer and lane down to the case, and has already captured what that number turned up. The other is still arguing about a fuel surcharge. One keeps the account at bid time. It won’t be the one holding a report it never acted on.

The Tool Was Never the Hard Part

The obvious fix is more software. It doesn’t work. You can run the same dashboards a billion-dollar shipper does. So can the company across town. What’s scarce now is the people who know which number to act on, and who go act on it.

Watch the companies that bought the tech and stopped there. PwC surveyed 767 operations leaders: 85% think they’re ahead on technology, 89% admit it hasn’t paid off. BCG says the difference comes down to how good your people are, not how new your software is. Gartner expects 60% of these digital efforts to fall short by 2028.

A dashboard shows you a number. It won’t tell you the inbound lane from your second-biggest supplier is padding its margin out of yours, and it won’t go fix that lane once you know. A problem you find but never capture isn’t savings. It’s a slide. That’s the case for keeping the finding and the fixing on one team. We staff for exactly that.

You Can’t Hire Your Way Out of This One

Then hire the people who can read the software. Good luck with that. The folks who model total cost of ownership and run network analytics don’t work at $50 million food companies. They’re at the giant carriers, the giant shippers, and the consultancies. The market made it that way, not you.

And it’s getting worse.

Over 76% of supply chain teams report staffing shortages. Hiring costs are up 22% in a year, demand for this talent is climbing 31%, and the role will grow five times faster than the average job through 2034. The rare ones with the scars and the analytics are getting 15-20% raises to jump ship. Not to you.

So build versus buy answers itself. You don’t build the department, and you sure don’t build both halves of it: the analysts who find the money and the operators who go take it back. You plug into one that already runs both. Ours took 28 years. A good surgeon doesn’t operate on himself. He finds the specialist down the hall. You should too.

Execution Goes Autonomous, and Half Measures Lose

While you’re deciding, the work itself is changing. Seeing your freight used to be enough. Now you have to act on it automatically. AI agents are starting to run the work, not just flag it, and that rewards the partner who finds the problem and fixes it in one motion.

Gartner expects 40% of business software to carry AI agents by the end of 2026, up from under 5% a year ago. McKinsey says this cuts 5-20% off distribution costs and halves forecast error. Mid-market companies move fastest through an outcome-as-a-service model because few can build an AI team in-house.

We already run this for our customers. Our work with Qued put AI into scheduling. It now confirms 95.5% of appointments untouched, 98% at the window we want. The shipper built none of it. They plugged in. The outfits that only sell reports, or only haul loads, won’t keep up.

The Cost of Waiting Just Got a Lot More Expensive

Add it all up, and waiting is the most expensive move you can make right now. Most deals I’ve lost didn’t go to a competitor. They went to the next quarter. “Let’s circle back after budget.” In a soft market, fine. The market isn’t soft anymore.

Truckload costs are running 16-17% above last year, with capacity tightening faster than anyone planned for. Service is harder to buy too. First-tender acceptance has slipped to about 85% from 92%, pushing more of your freight into a spot market sitting 25% higher. Ship refrigerated? Reefer rates are getting squeezed hardest as capacity dries up.

And the analysis commits you to nothing. It only shows you what to fix next. Most of what we find is hiding in a network the customer swore was already tight. The finding is the cheap part. The waiting is what costs you, because the fix gets pricier every month the market tightens while the leak you haven’t caught keeps running. Give it another year, and the bleed hardens into next year’s baseline. It won’t pause for planning season.

Standing Still Is a Side of the Line

Two years from now, the distance between the shippers who moved and the ones who waited will show up in their margins, and in who’s still holding the account. That distance is cheap to close today. It won’t be later.

Take one food manufacturer we worked with. We went at their total landed cost, the full delivered cost of their product, including inbound freight, accessorials, and the freight buried in supplier pricing, and found $600,000 in net savings with another $1.25 million sitting on the table.

Finding it is half the job. We go capture it too. That’s the bench a $50 million shipper can’t build alone, and you don’t have to.

I’ve watched sharp operators wait a quarter too long, mostly because nobody put the number in front of them until it had already hardened into the cost of doing business. I’d rather you saw it now, while it’s cheap to fix. Don’t operate on yourself.

See what your network’s costing you before it becomes next year’s baseline. Talk to NT Logistics.