PITT OHIO Rebuilt How 2,100 Drivers Work and Saved $11 Million in Year One. Then It Pointed AI at the Inbox.

THE CRAFT

In March 2004, Chuck Hammel told a Pittsburgh business magazine something most trucking executives still don’t believe: “It’s the information flow that goes along with it that’s actually more important than the freight movement itself.” His family’s less-than-truckload carrier, PITT OHIO, was spending $8 million a year on information technology at the time — a family freight company, twenty-two years ago, putting real money behind the idea that the trucks were the easy part.

I want you to hold that quote against the year we’re all living through. Right now, somewhere in your inbox, there is a pitch for an AI deployment. The board has asked about your AI strategy. The pressure is to buy something — a copilot, an agent, a platform — point it at the business, and report progress. And the quiet problem with most of those purchases is arithmetic nobody runs: AI is a multiplier, and a multiplier is only as good as the number underneath it. Multiply a workflow you understand and you compound. Multiply a mess and you get faster mess.

PITT OHIO is what the other path looks like, run patiently, for two decades, by one family. Between 2023 and 2025 the company rebuilt how its 2,100 drivers work — and documented more than $11 million of savings in the first year — and then, only then, pointed AI at the least glamorous target imaginable: the inbox where customers email pickup requests. Requests that took a person up to fifteen minutes each now become orders in the company’s freight system in seconds. Nobody got a press tour of a robot. The mail just started answering itself.

The trade press has covered the pieces — an award here, a vendor announcement there. What nobody has written is the part your business can actually use: why it worked. The answer isn’t in the 2025 deployment. It’s in the 2004 quote, and in everything the Hammels did between the two. That order of operations — the twenty unglamorous years before the multiplier — is what this brief is about.

THE OPERATOR

The situation

Less-than-truckload freight is a business where the margin lives in minutes. A carrier like PITT OHIO consolidates palletized shipments from many customers onto shared trucks, runs them through a network of terminals, and delivers them across the Mid-Atlantic and Midwest. The freight itself is commodity. What separates a carrier that earns money from one that dies is operational: how fast a dock turns, how tight the line-haul network runs, how many minutes of paid human time each shipment consumes on its way through the system. The industry keeps score in a brutally simple way: how many cents does it cost to earn each dollar of revenue? The best carriers spend under 80 cents; the strugglers spend nearly the whole dollar.

The past three years have been a proving ground for that arithmetic. The freight recession that began in 2022 ran longer than any downturn in a decade. Yellow Corp — the country’s third-largest LTL carrier, holding an estimated 10 to 15 percent of the market — collapsed into bankruptcy in August 2023. And the pain hasn’t finished working through the survivors: in the first quarter of 2026, ABF Freight, the closest big public comparison to PITT OHIO by revenue, was spending more than 97 cents of every dollar it took in just to run the business — leaving under three cents of profit on the dollar, worse than the year before — and announced it would close ten terminals and cut 2 percent of its workforce. Old Dominion, the industry’s best performer, watched tonnage fall 7.7 percent. This is the market a mid-sized family carrier has been operating in.

PITT OHIO has seen worse. Charles Hammel Sr. started hauling for J&L Steel in 1919 with a horse and a buggy. His son incorporated the business in 1960. Chuck Hammel III founded PITT OHIO in 1979 with his brothers Bob and Ken — three trucks and a leased warehouse in East Liverpool, Ohio — and almost immediately lived through the industry’s formative trauma: deregulation. “We had operating authority for just the state of Pennsylvania that was worth $1 million one day, (then) worth zero the next day,” he told Smart Business in that same 2004 interview. Read that quote carefully, because I think it explains the next forty years: the Hammels learned early that the protected asset is never the moat. The operating discipline is. Chuck bought out his brother in 2003 and has been sole owner since.

And here is what the work actually looked like inside that discipline, before the recent rebuild. Drivers keyed manifests by hand — the company’s own case study later measured a 75 percent reduction in manifest-entry time, which tells you how much typing there was to remove. Dispatch, driver workflow, line-haul planning, and maintenance requests lived in separate systems. And in customer service, a request as simple as “come pick up three pallets on Thursday” arrived as an unstructured email, and a person spent five to fifteen minutes reading it, interpreting it, and re-keying it into the freight system. Hundreds of those arrive every day. None of this was broken, exactly. It was ordinary. Which is precisely why the margin was hiding in it.

The move

The rebuild came in two layers, in an order worth writing down.

The first layer was the workflow itself — no AI involved in the way you’d recognize the word. Starting in 2023, PITT OHIO replaced its patchwork of dispatch, driver, and telematics systems with a single operations platform, built with an outside trucking-technology partner. Pickup-and-delivery dispatching, the driver’s mobile workflow, line-haul planning, electronic logging, and maintenance requests moved onto one spine, so that what a dispatcher sees, what a driver taps, and what the shop learns about a truck stopped being three separate conversations. The company put 2,644 people onto that platform — 2,127 of them drivers — across 25 operations centers, in nine months. For a family carrier doing this mid-recession, that speed is the tell that leadership treated it as an operating decision, not an IT project.

The savings arithmetic is the part I’d tape to the wall. The company documented $11,349,000 in first-year cost savings — and the overwhelming bulk of it, about $11 million, came from one source: roughly twenty minutes of recovered time per driver per day, multiplied across 2,100-plus drivers. Not a moonshot. Not a new revenue line. Twenty minutes, across the whole fleet, compounding every working day. SupplyChainBrain, reporting independently on the deployment, put the figure at “more than $11 million” and added the detail that tells you it was real: in 2024, PITT OHIO absorbed over a thousand new shipments a day — including a surge of freight from a failed competitor — without adding drivers or back-office staff. Against roughly $950 million of revenue, $11 million means the rebuild handed back a little over a penny on every dollar the company earns — in an industry where the winners and the strugglers are separated by pennies.

The second layer is the one that made the AI press releases, and it started from a confession. Scott Sullivan — PITT OHIO’s EVP and CIO, with the company since 2001 — has said plainly that before this project, “outside of searching ChatGPT, Pitt Ohio had never worked with AI.” In 2023, with the workflow spine going in, his team picked a first AI target that would embarrass a vendor’s demo reel: the pickup-request inbox. His team built N@TE — “No-Touch Email” — wired by API into the freight system PITT OHIO already ran. A filter separates ordinary correspondence from pickup requests; N@TE reads the unstructured email, extracts the shipper, receiver, commodity, dimensions, and special instructions, and creates the pickup order — as Sullivan puts it, “The response time back to the customer now is pretty much immediate.” It launched in June 2025. The customer-service people it helps were not shown the door; Sullivan again: “A customer service job is more than inputting pickups into a system.”

Notice the architecture, because it’s the part a leader can copy on any budget: nothing was ripped out. The freight system stayed exactly where it was, as the system of record. The AI layered on top as the system of action, finishing one whole workflow — email to order — under human supervision. The industry will sell you this pattern under many names; what matters is that a family trucking company in Pittsburgh executed it against its most boring workflow first, and got paid immediately.

And none of this was a first date. This spring, N@TE won PITT OHIO its sixth CIO 100 award — the same technology honor it took home in 2007, 2010, 2013, and 2017 for work that had nothing to do with AI, and in 2024 for the operations platform. The company holds a real patent portfolio: nine granted, from terminal microgrids that power its docks with wind and solar to machine-learning systems for tracking and cutting carbon emissions — and the named inventors include Chuck Hammel IV. The fifth generation of this family files patents. PITT OHIO holds a U.S. patent on N@TE itself and licenses the system to the rest of the trucking industry — the hundred-year-old freight company, quietly, becoming the technology seller. As a kicker to a hundred-year story, it’s hard to beat.

The result

Here is what the record supports, with the sourcing stated plainly — because for a private company, who says a number matters as much as the number.

The results land the way an operator feels them. The driver-workflow rebuild returned more than $11 million in its first year — the arithmetic is almost boring: about twenty minutes back per driver, per day, across 2,100-plus drivers, every working day of the year. Manifest entry got 75 percent faster. In 2024, PITT OHIO took on more than a thousand additional shipments a day and moved them with the drivers and the back-office staff it already had. And a pickup email that used to eat five to fifteen minutes of a person’s time now becomes an order in seconds. Those are the numbers I’d put my name to; the flashier ones the company markets around N@TE, I’d leave on the table — the four above already tell the story, and they’re the kind a peer can check against his own operation.

The business kept growing through the worst freight market in a decade. PITT OHIO self-reported $828 million of revenue in 2020 and “over $970 million” for 2025; Transport Topics independently pegs the broader PITT OHIO Transportation Group at $1.02 billion for 2025, ranking it the 44th-largest for-hire carrier in the country. While peers retrenched, the family bought: a former Yellow terminal in Milwaukee for $19.3 million in January 2024, and Sutton Transport — a 400-truck Midwest carrier — effective January 2025, a move projected to make the group the 12th-largest LTL operation in America. A private company can keep its margins to itself. It cannot hide expansion.

Set that against the same quarter’s public peers: ABF down to under three cents of profit on the dollar and closing terminals; Saia’s costs creeping the wrong way as it digests growth; Old Dominion holding its famous discipline but shrinking. The freight cycle is finally turning — the ATA’s tonnage index just posted its best quarter since 2017 — but the recession sorted the industry into companies that spent the downturn cutting and companies that spent it building. PITT OHIO built.

Now the part I’d want you skeptical about, because a brief that only flatters is an advertisement. Some of PITT OHIO’s growth rode the tide, not the technology: when Yellow collapsed, its freight and its terminals landed on the survivors, and the freight market has been turning back up all year. The rebuild sharpened a company that was already going to grow. It’s also a private, family-owned business, so the full margin picture stays inside the family — you’re taking the shape of the story on the strength of what a peer can independently see: the volume it absorbed, the awards, the acquisitions, the twenty-year spending pattern. And the succession is real: in June 2026, after forty-six years as President, Chuck Hammel stepped back to Executive Chairman. The discipline he built now has to run without him at the controls. None of that breaks the story. It’s the honest frame around it.

The Craft of AI read

Strip the trucking specifics away and here is what I think a leader running a $100 million or a $2 billion company should take from Pittsburgh. Experience is the advantage; AI is the multiplier — and PITT OHIO is the cleanest demonstration I’ve found this year of what the multiplier does when there’s actually something underneath it. The 2004 quote wasn’t a slogan. It was an operating conviction the family paid for, in cash, for two decades — $8 million a year of IT spend when that was real money for a regional carrier, six technology awards across nineteen years, a patent portfolio a fifth-generation Hammel contributes to. So when the AI wave arrived in 2023, PITT OHIO didn’t have to begin by understanding its own workflows under pressure, with a vendor’s meter running. The understanding was already on the shelf. The AI had a number to multiply.

Compare that with the sequence being sold to you right now. Almost every AI sales pitch leads with something to buy — an agent, a set of copilots, some “software.ai” platform — and it lands on your desk before anyone has understood, at the level of the people actually doing the work, how that work happens. You’re handed the tech solution first, and the understanding is supposed to catch up later, if it ever does. But if nobody in your business can describe, at ground truth, how a pickup request becomes an order, how a claim becomes a check, how a quote becomes a booking — every step, every re-key, every wait — then the AI has nothing real to work on. Sullivan’s team could point AI at the inbox because the inbox’s job was already understood and the redesigned workflow spine gave the output somewhere clean to land. The boring first target wasn’t timidity. It was precision.

Notice, too, what they did not do. They didn’t rip out the freight system and start over; it stayed as the system of record while the AI layered on top as the system of action — proof, at family-company scale, that you can leave what works and layer what’s missing. And because the workflow understanding was theirs, the resulting intellectual property was theirs too. When the vendor is the only one who understands your workflow, the vendor owns the future. When you are, you do — sometimes literally, in patent form.

The transferable lesson is not “spend twenty years.” You don’t have that, and you don’t need it. The lesson is the order: understand one workflow at ground truth, redesign it, then let AI finish the routine version of it under supervision — and start with a target so unglamorous nobody will defend it. PITT OHIO’s was a shared inbox. Yours is probably something equally invisible, absorbing operating margin today in fifteen-minute increments no one has ever added up.

Things to consider

• A multiplier needs a multiplicand. Before you approve the next AI line item, ask what it will be multiplying. If no one can walk you through the target workflow as it actually runs — every system, every handoff, every re-key — the honest answer is “nothing yet,” and the first money should buy understanding, not software. PITT OHIO’s AI worked in months because the understanding was two decades deep.
• The margin hides in minutes, not moonshots. Eleven of PITT OHIO’s first-year millions came from twenty minutes per driver per day. Run that arithmetic on your own business: pick one role, estimate the minutes lost daily to typing, chasing, and re-keying, multiply by headcount and loaded cost. The number is almost always larger than the AI pilot you were about to fund, and it is being spent today, this quarter, whether you look or not.
• Pick a beachhead nobody will defend. The pickup inbox had no constituency — no one’s identity was invested in re-typing emails. That’s what made it fast: no politics, immediate payback, and a workforce that experienced AI as relief rather than threat. The showcase project with nine stakeholders is where first deployments go to die. The boring one is where they compound.
• Decide who owns the understanding before the vendor arrives. PITT OHIO built with small partners, but on its own workflow knowledge — and kept the intellectual property, down to a licensing deal. Reverse that dependency and you rent your own operating model back. Before any AI statement of work is signed, settle in writing whose asset the resulting workflow design is.
• A practice outlives a project. Six technology awards across nineteen years is not a project; it’s an institutional habit — one strong enough that the founder just stepped back to Executive Chairman without the machine pausing. Ask what your version of the habit is: who examines a workflow in your company when there’s no initiative, no consultant, and no board question forcing it?

THE WORKBENCH

Do this tomorrow

This one requires an hour, a notepad, and one shared inbox.

Find your pickup-request inbox. Somewhere in your business, unstructured demand — emails, PDFs, voicemails, portal messages — arrives and a paid person translates it into a structured system entry. Order intake, claims setup, quote requests, patient referrals, load tenders: the shape is identical. Go sit with the person who does it. Count how many arrive in a day, time three of them start to finish, and note every system they touch. Then do the PITT OHIO arithmetic on your own numbers: items per day, times minutes each, times the people doing it, times loaded cost per hour, times 250 working days. Write the annual figure down. That is operating margin being absorbed, today, by translation work nobody has ever mapped — and in most mid-market companies it’s a six-figure number hiding behind a task everyone considers too small to examine.

Then ask two questions, in this order. First: if this intake ran the way it would be designed from scratch — the request arriving, being read, and becoming a clean system entry with no human re-keying — what would remain for a person to do? (At PITT OHIO the answer wasn’t “nothing”; it was the judgment calls, which is why the customer-service team is still there.) Second, and only second: is this workflow understood well enough, right now, that an AI could finish its routine version under supervision — or did the first question just reveal that nobody actually knows how it runs? If it’s the latter, you’ve found your real starting point, and it costs a notepad, not a license.

The rigorous version

That exercise finds one intake point. The rigorous version does that work across the business, with the people who already know where the truth lives, and leaves you with a design instead of a diagnosis.

Every AI strategy you’ve been sold starts with vendors. This one starts with the small group of people in your business who already know where the operational truth lives. Two days in person. A target operating model that has AI built in from the ground up. A 90-day starting plan.

You walk out with a ground-truth map of how your core workflows actually run, the two or three where redesign releases real operating margin, a target operating model with AI placed where it multiplies something genuine, and a 90-day starting plan your own team can execute. Price: $20,000. I run a small number of these each quarter.

THE QUESTION

Chuck Hammel spent forty-six years as President of the company he founded with three trucks, and in June — the same season his technology practice collected its sixth award in nineteen years — he stepped back to Executive Chairman. The thing he built that will outlast his tenure isn’t the terminal network or even the patents. It’s a habit: the institutional reflex of examining how the work actually happens, decade after decade, whether or not anything called AI exists. The 2004 version of Hammel had no idea a language model would someday read his customers’ emails. He just believed the information flow mattered more than the freight — and kept paying for that belief until it became the most valuable thing the company owned.

So here is the question, and I’d ask you to answer it in under thirty seconds, honestly: where does unstructured demand enter your business — and can you name, right now, the number of minutes a paid person spends translating each piece of it into your systems? Hammel’s people could name it: five to fifteen minutes, hundreds of times a day. That’s why their AI worked. If you can’t name yours, the gap between you and Pittsburgh isn’t a technology gap. It’s a looking gap — and it’s costing you margin this quarter, not someday.

If you want to talk through what the rigorous version looks like for your company — the two-day workshop, the target operating model, the 90-day starting plan — hit reply, or send a note to grant@thecraftofai.com. I read every one, and the answers are usually where the next brief starts.

— Grant K. Baldwin grant@thecraftofai.com

---

Want to do this to a workflow of your own?

I’m an investor in geniant, so you know where I stand. For more than 25 years, the people behind it have done one thing: understand how work actually happens, then build software that works the way humans do. That’s exactly what this brief describes — they sit with how your work really happens, redesign the one workflow that’s quietly bleeding margin, and layer AI on top, shipped to production by one senior-led team, your systems of record left in place. In weeks, not months.

Book a discovery call →