Back
Why the Future Enterprise Won’t Hire More Software. It Will Build a Digital Workforce.
Robert Nathan
I’d bet a month of margin there’s a software demo on your calendar inside the next 2 weeks. 45 minutes, a sales engineer driving, a deck with the word orchestration on slide 3.
Cancel it.
I sell technology to freight brokers for a living, and I’m telling you to cancel it because that demo is application number 306, and 306 will cover exactly as many loads as the first 305 did. I earned this opinion the expensive way. I ran freight, I built transportation tech, and I’ve put money into logistics companies as an investor, and from every one of those chairs I watched the same reflex: volume outruns the floor, finance kills the reqs, somebody buys software to split the difference. And through all of it, I paid humans to send emails. Still can’t believe it.
My position after 20 years of signing both kinds of paychecks: enterprises won’t scale by adding applications. They’ll scale by deploying autonomous digital teammates that execute work inside the systems they already own. We build one named Ellie, and 75% of freight in our live accounts moves through her.
But park the product for now. The argument stands even without it.
You’ve Run This Experiment 305 Times Already
Go count your logins. All of them, including the vetting portal nobody admits to and the spreadsheet with a version number in its name that 3 of your reps treat as the actual system of record.
Zylo counted for the rest of the market: 305 SaaS applications at the average company, with spend up 8% on a portfolio that’s stopped growing. 305 separate attempts to scale without hiring, and your load’s still uncovered at 4PM, because the work never lived inside any one of them.
Your reps pay the tax daily. HBR followed 137 workers at 3 Fortune 500 companies and clocked 1,200 app toggles a day, almost 4 hours a week lost to finding their place again.
FreightWaves has been writing brokerage obituaries all cycle. Not one of the deceased died short on software.
Not One of the 305 Ever Covered a Load
I open every sales call the same way. How does your carrier team source trucks? I’ve asked it hundreds of times and gotten a single answer: pull the load details, hunt capacity, write the emails, work the phones, verify compliance, chase confirmations until 6.
Now find the software in that list. It’s not there. The TMS shows up after the fight’s over to write down who won.
Gartner now describes agentic AI as a virtual workforce that moves past insight into execution, which is the politest available way of admitting 20 years of enterprise software watched from the bleachers. The vendors heard the whistle too, which is why 40% of enterprise applications will ship with task-specific agents by year’s end, up from under 5%. When a software company staples a worker onto its own product, it’s confessing what you were actually paying for.
Your copilot pilot is little more than a suggestion box that handed your most expensive rep the mouse.
A Macro, a Temp, and a Teammate Walk Onto Your Floor
Since the whole category answers to “AI,” let me split it the way an operator would.
The macro is automation. Rules, scripts, RPA. It fires the email at 8:02 sharp and falls over the second the portal redesigns its login page or the commodity doesn’t fit the template. Useful, cheap, and it does steps. But steps were never the job.
The temp is an agent. Smarter, handles a real task, drafts a decent carrier email. However, it lives inside 1 app, works 1 session at a time, remembers nothing, and needs a babysitter. Nobody ever scaled a floor on temps.
The teammate is a digital workforce, and the word that separates it is ownership. It takes the outcome, covers the load, and works every system that outcome touches: TMS, load boards, portals, inbox. You manage it the way you manage people. Set the guardrails, review the output, hold it to a number.
Automation does steps, an agent does a task, and a workforce owns the load from first email to signed rate con.
Read the Comp Plan You Signed for Your Software
A per-seat vendor’s revenue is your headcount. More humans logged in, bigger renewal. You hired a company whose paycheck grows every year you fail to get leaner, then wondered why the efficiency never showed up. I sold under those rules. The model deserved to crumble.
And crumbling it is. Pure per-seat pricing fell from 21% of vendors to 15% in 1 year while hybrid models jumped from 27% to 41%, because a teammate who executes the work deletes the seat she was billed under. A digital workforce prices like labor, against the cost of the work, and that’s the frame I’d take into any buying decision this year.
Freight’s already picking sides. Truckstop and Bloomberg Intelligence found 41% of brokers deploying AI while 48% sit out, with more of them betting on margin than revenue this year. Services and SaaS are going to converge into LaaS. Chairs don’t survive that trip.
The Amnesia Problem
Every operator has hired this guy at least once. Great first day. Picks up your lanes fast, even clocks which carrier answers on the second ring. Then Tuesday morning he asks where the coffee machine is, because as far as he’s concerned, Monday never happened.
Most enterprise AI is that guy, and you signed him to a multi-year deal. MIT’s Project NANDA found 95% of enterprise AI pilots produced no measurable P&L impact, and the dead ones shared his defect. They couldn’t remember anything. Ask your reps when they stopped opening the last pilot. My guess is somewhere around the 400th time somebody typed the same customer’s dock rules into a blank box.
Half the definition of a digital workforce is that it remembers. Every negotiation and every booked load compounds into context the operation owns. Ours is called the Carrier Context Graph, and that context matters more than the model underneath it. It’s the only thing you’ll ever pay for that gets cheaper the longer it works.
What Your Best Rep Does Instead
Memory is what makes the teammate worth keeping. What it changes is the job description underneath, because once the digital workforce owns the repetitive execution, the humans move up the value chain. Nobody’s floor empties out. The morning just starts finished, and your rep walks in at 7 to a different kind of day.
Exceptions Over Execution: Outreach ran overnight and compliance cleared early, so what’s waiting is a short queue of the loads that need a person. Your rep’s first hour is decisions, not data entry.
Relationships Get the Hours Back: The carrier you’ve worked with for 10 years wants a voice on the phone, not a templated email at 8:02. That call is where margin and loyalty live, and it’s the first thing that falls apart when a rep is buried in clicks.
Judgment Becomes the Job: The rate that smells wrong, the shipper SOP nobody documented, the driver who’s 3 hours out and quiet. Machines escalate those; people settle them.
The Role Moves From Task Executor to Outcome Owner: Your rep stops being measured on activity and starts owning the rate, the carrier, and the customer. It’s the strongest retention move on the board, because the reps worth keeping are the ones most bored by the clicking.
The Pyramid Reshapes Under Them: 55% of supply chain leaders told Gartner they expect agentic AI to cut entry-level hiring, and Staffing Industry Analysts now covers digital workforce providers the way it covers staffing firms. The reps who stay are learning to lead machines, and every generation of business gets a new leadership discipline. This is ours.
Cancel Their Demo. Book Her Interview.
Now un-park the product, because an argument this size needs proof in production and I’ve got some. Ellie is the digital workforce running today: the execution layer working your existing TMS, portals, load boards, and inbox, with TOAS showing leadership every action live instead of in Friday’s report. Michael Cherney at Cooler Logistics said it cleaner than I can: new level of scale, same headcount, reps off the busywork.
So interview her like a hire, not a vendor. Book 30 minutes and refuse the sandbox. Bring your live portal and your ugliest lane, and type the ask the way your rep actually talks: need 3 reefers out of Laredo to Dallas Monday morning, around $2,100. If she stalls in your workflow, you stumped the AI guys, and it cost you half an hour.
Check her references. Then email me, robby@tryenvoy.ai, with the question every candidate in this category should answer: what percentage of your customers’ freight books through the tool today? I’ll answer in 1 line. Ask everyone else courting you and time the pause.
You’re running 305 applications and 0 teammates. Fix the second number.
