Beyond SaaS: When Technology Becomes Part of the Workforce

Let’s argue about whether AI belongs on payroll after you explain why you’re hiring someone to do a job it can already handle. I spent years paying good carrier reps to keep expensive software useful, and I’d love to get some of that money back.

Ellie, our agent and execution layer at Envoy, negotiates with carriers inside your max pay and brings the booking back for your rep’s approval. Roughly 75% of freight in our live accounts is booked through her. If she showed up with a résumé and asked about dental, you’d know exactly where to put her in the budget.

Going beyond SaaS means putting Ellie on payroll beside your reps and expecting those reps to cover more freight with her. If your 2027 plan needs 10% more people to move 10% more freight after hiring Ellie, you should ask me what I sold you in the first place.

Is SaaS Dying, or Is Wall Street Just Being Dramatic?

Yes, SaaS is dying, and Wall Street was late to the funeral. There’s no future in selling software to freight brokers, and I say that as a guy who sells to them for a living.

The seat model has always been stacked against you. Your vendors bill by the login, so every rep you hire gives them a raise. They cheer your reqs louder than your recruiter does. So how hard do you think they’re working on a tool that cuts headcount?

Investors quickly caught on, and software stocks lost more than $1 trillion in the first five weeks of 2026. Forbes pinned it partly on AI tools that cut how many licenses companies need.

Then Workday, the HR giant that bills per employee, called itself the platform for “people, money, and agents” in its February earnings release. When the headcount company starts counting agents, take the hint.

Does AI Belong in Your IT Budget or on Your Payroll?

If AI is doing the work, put it on payroll. If it’s helping a person do the work, put it in IT. This really isn’t complicated.

Ellie sources carriers, communicates, negotiates, verifies, and escalates. Those are jobs your ops team gets paid to do. Calling her “software” because she logs into a TMS is like calling your rep a laptop expense.

The reason companies shove AI into IT is obvious: software can hide in a pilot. Labor has to produce.

So stop giving AI some cute innovation budget where everybody gets to clap for the demo, and nobody has to defend the economics. Put Ellie next to the reps you planned to hire and judge her the same way you’d judge them: how much work got done, what it cost, and how many additional people you no longer need to add as volume grows.

Gartner found CFOs expect headcount growth to fall from 6% to 2% in 2026. That shift is already happening. The question for freight CFOs is whether they’re going to budget AI based on the invoice it sends them or the payroll it replaces.

Should You Be Hiring Carrier Reps Right Now?

No. At least not until you’ve proven the work requires another human.

Freight has a terrible muscle memory here. Capacity tightens, phones get busier, somebody opens a req. We did this at scale in 2021, then spent 2023 laying off the people we swore we desperately needed. Apparently “we overhired” sounds better when HR calls it rightsizing.

Now capacity is tightening again. Spot rates were near $3.38 per mile and tender rejections around 14.4% in early September. At the same time, 86% of brokers said capacity was getting harder to find, 43% said margins were shrinking, and 53% still planned to hire more brokers.

Read those numbers again. Margins are getting squeezed, so the response is more payroll?

When we launched Ellie Workforce, Cooler Logistics CEO Michael Cherney said his team reached new scale without adding headcount. That should be the hurdle now. Before you open another carrier rep req, make your existing workforce, human and AI, prove it can’t absorb the volume.

Isn’t Offshoring the Carrier Desk the Cheaper Answer?

Only if you’re still solving for the price of a human seat.

Offshoring made perfect sense when the choice was a carrier rep in Chicago or a cheaper one in Manila. Plenty of brokerages took that deal, shaved labor costs, moved some repetitive work overseas, and called it transformation. Legit. The problem is the model never changed. Every load still needed a person touching it, so growth still came with a hiring plan attached.

Now the question is different. How much of that work needs a person at all?

Even BPOs see where this is going. Portage’s 2026 report says providers are moving away from seat-based work and toward AI-enabled workflows. That’s why Ellie can be white-labeled by BPOs that would rather sell completed work than bargain bin headcount.

Shippers are making the same calculation inside their own load control centers. If your edge is still “our humans cost less,” you’re bringing a labor-arbitrage pitch to a workforce-automation fight.

Who Gets Sued When the AI Books a Bad Carrier?

You get sued. Of course you do. Ellie isn’t going to sit for a deposition while your brokerage watches from the parking lot.

If Montgomery v. Caribe still feels like breaking news to you, you have some catching up to do. The Supreme Court already made clear that negligent-selection claims against brokers can survive federal preemption, so the carrier you put under the load is very much your problem.

What AI changes is how much improvising happens before that carrier gets there.

A rep with a pile of open loads, a shipper breathing down his neck, and a truck sitting 20 minutes away can convince himself a warning probably isn’t a big deal. Humans are fantastic at finding religion after the accident.

Ellie runs the same checks through Highway and MyCarrierPortal every time, and TOAS keeps the record.

If you’re getting sued either way, if I were you, I’d rather explain the process with evidence and timestamps than explain why somebody skipped it.

What Happens to Your Reps When Ellie Does the Work?

They stop spending half their day proving they’re busy.

I’m not going to pretend that the carrier rep job looks the same five years from now. It won’t. If somebody’s value comes from making 80 calls, copying appointment times, and chasing the same carrier for the third update, Ellie is going to eat a lot of that job.

Good. That was never the part worth paying humans for.

Keep the rep who knows when a carrier going conditional is a real problem and when it’s noise. Keep the one who knows when to break the rate guardrail, when a relationship deserves a phone call, and when something smells wrong before the system can explain why.

That’s what becoming bosses of machines looks like in practice. Your reps stop babysitting every routine move and start stepping in where their judgment is worth something.

So Send Me the Headcount Plan First

Before you approve another carrier rep req, send me the list. Seriously.

You’ve spent this whole article hearing me argue that AI belongs on payroll, another hiring spree probably isn’t the answer, offshoring only makes the old labor model cheaper, and your best reps should be making decisions instead of babysitting loads. So if your 2027 plan still says volume up 30%, headcount up 30%, I want to see why.

Ellie runs through a Chrome extension on your TMS and already books roughly 75% of the freight across our live accounts. The work also builds your Carrier Context Graph, so the operation gets smarter instead of simply getting bigger.

We hired our way through the last hot market and spent the soft one undoing it. I have zero interest in watching freight run that stupid play again.

Book 30 minutes, bring your headcount plan and your ugliest lane, and let’s see which reqs survive once Ellie works the freight. Or email robby@tryenvoy.ai your planned hires and loads per rep. I answer those myself.