The Things You're About to Do That You Never Imagined

There is a particular kind of conversation I keep having lately. Usually with another operator. Usually after the third coffee. Usually about something that started as small talk and became a confession.

It goes like this:

I am building something I never thought I would build. I am thinking about the future of my business in ways I never thought I would think. I am considering identities and roles and possibilities that, three years ago, I would have laughed at out loud.

I would have laughed too. I am no longer laughing.

What follows is the list of things I believe are coming. Some are predictions. Some are bets I am already placing with real capital. All of them are conclusions I have arrived at after twenty-six years in freight, twenty-four months of building AI tools at the edge of what is possible, and a quiet, growing certainty that my industry is about to look nothing like itself.

If even three of them are right, the world looks very different. I think more than three are right.

You do not get to opt out of this one. The shipper expects API integration, real-time visibility, document automation, and instant quoting. The carrier expects digital settlement, electronic POD, and one-click rate confirmation. Customer service expectations are now set by every B2B SaaS company in any other industry — not by your local competitor.

If you are not building software, or buying it intelligently and integrating it deeply, you are competing in a different sport than the one you signed up for. Most brokers will admit this only after they have lost the customers they thought they could not lose. A few will admit it in time.

Too small to absorb consolidation. Too big to move fast. Too dependent on traditional dispatcher labor to compete on margin with the AI-augmented solo operator. Too lacking in personal brand to compete on relationships with the one-or-two-person operations.

The 1–2 person AI-augmented operator runs at premium margin. The 500-employee enterprise broker runs at scale and absorbs the consolidation deals. The 50-employee shop runs at neither. Watch for them to either shrink down to five or ten people with massive AI leverage, get acquired by someone bigger, or quietly close.

This is not me being mean. It is me reading the math.

This one already shows up in the early data. A solo operator with three or four well-configured voice agents, a custom quoting layer, and a cleanly integrated TMS can quote faster, cover lanes faster, dispute carriers more cleanly, and close more loads per rep than a five-rep team running on legacy systems.

The output is roughly the same. The cost basis is wildly different. Run the math forward four years and the solo shop is closing at thirty percent margin while the fifty-person shop is closing at single digits and wondering why their best reps keep quitting.

Capital and headcount used to be the moat. Now they are the burden.

Watch this one carefully. We are starting to see early signals — a few acqui-hires in adjacent industries, a couple of strategic minority stakes inside our own. The interesting freight deals over the next twenty-four months will not be "C.H. Robinson buys a fifty-million-dollar brokerage." They will be "Schneider buys a three-person AI-native broker for the operating system."

That is the deal nobody is structuring yet. Within twenty-four months, everybody will be. If you are building a tight, modern, AI-native ops stack, you are not just running a small brokerage. You are running a software company with a freight P&L attached — and the asset brokers are going to want what is inside the laptop more than what is on the books.

The whole supply chain is shifting one position to the right. AI is dissolving the gatekeeping that used to separate the roles. A dispatcher with the right tools can quote, post, and book — that is the broker's job description. A driver with the right tools can dispatch his own runs and find his own loads — that is the dispatcher's job description. A shipper with the right AI agents can post, vet carriers, manage compliance, and handle exceptions — and on commodity lanes, that means the broker is no longer required.

The brokers who matter in five years will be the ones who bring something the shipper, dispatcher, and driver collectively cannot do for themselves: judgment, relationship, dispute leverage, capital position, and unique data on hard lanes. Everything else gets automated out of the equation.

The freight broker with twelve thousand LinkedIn followers, a YouTube channel, and a written record of opinions will quietly out-earn the freight broker with eighty employees and no public presence. This sounds wrong to most of my industry. Most of my industry has not yet noticed that the customers paying premium rates in 2026 heard about you from a podcast, not from a cold call.

Being known is now a more reliable form of compounding than being big. Being trusted at internet scale is harder than being trusted at local scale. The reward, when it comes, is also bigger.

Twenty-four months ago, the line between "person who runs a freight company" and "person who builds software" was a solid wall. You were on one side or the other. Today the wall is a dotted line. Within twenty-four more months, it will not exist as a meaningful distinction at all.

The operators who refuse to learn how to build will be priced out by the operators who do. The builders who refuse to learn an industry deeply will be out-domain-knowledged by the operators who built the same tools faster because they actually knew the problem. The middle, where you are neither, is the worst place to be.

Cross over. It is not as scary as it looks. It is significantly scarier to stay where you are.

I am writing this from a desk where, in the last seven days, I have deployed three voice agents into production, replied to a twenty-lane RFQ in forty-eight minutes using a quoting skill I built myself, killed an eight-hundred-dollar-a-month software subscription, and seriously considered hiring my first international developer.

Five years ago I was running twelve to fifteen browser tabs and praying. The version of me that walked into Jim Evans's office in December 1999 would not understand any sentence in this paragraph.

You are about to be that version of yourself for someone else. You are going to do things in the next thirty-six months that your past self would have called impossible. The only question is whether you go willingly or you get dragged.

Go willingly.

These are not predictions I make casually. I have skin in every one of them. Some I have already bet on with capital. Some I am betting on right now, this quarter, while you are reading this.

I do not need you to agree. I need you to think about three things:

One: Which of these would you have laughed at three years ago, and what changed?

Two: Which of these are you quietly already doing without admitting it out loud?

Three: Which of these are you most afraid to be wrong about?

You are one of those people. You may not have admitted it yet. That is fine. Get back to me in eighteen months.