Everything you touched today arrived at your fingertips through the global supply chain: the coffee at your favorite neighborhood cafe, the iPhone in your hand, the cardboard box carrying your latest impulse purchase on your porch.
None of it got there by magic. And many of the items nearly didn’t arrive.
They get to their destinations because somewhere in middle America, a warehouse supervisor is closing gaps by hand — a shift coming up a dozen people short, the count in the system not matching what’s on the shelf, or a line stalling with a truck already idling at the door — before the day goes sideways.
That scramble is the reality of the supply chain. The physical economy runs on it, and even though nearly everybody depends on it, most people rarely see it.
Over the last thirty years, software drove the cost of information, communication, and now intelligence itself toward zero. That’s why the digital world feels abundant. But the physical world never got that treatment. Things still cost too much and take too long to arrive, and the chain is coordinated by overloaded people operating based on their gut at the crack of dawn. The part of the economy everyone actually depends on is the part software never reached.
We know, because we’ve spent the past five years inside it.
Five years on the warehouse floor
Since 2021, we’ve been building the best and fastest growing tech-enabled staffing platform in America. We’ve filled over a million shifts, half of which were in the past year alone. Getting consistent, reliable workers to the right facility, at the right time, is arguably the hardest problem to solve in the supply chain. Delivering on that promise has meant embedding in the day-to-day of 3PLs, manufacturers, and distribution centers, living and breathing their operations.
We know their pain because we solve it every day. We’ve sourced hundreds of workers within a few days notice, backfilled shifts delayed by line outages no one flagged until it was too late, and helped customers sharpen labor decisions that had been made on gut feel and last year’s outdated projections.
The trust we built with the best industrial businesses in the country gave us the right to solve pain points outside labor. Workforce software can see the idealized plan but not the gaps being filled by hand. Labor marketplaces can deploy workers but stay blind to the decisions being made upstream. By embedding in our customers’ operations, we do both: We own the supply of labor and the demand for it, along with the ground truth of what actually happens when the two meet, across thousands of sites and employers. That position is earned one morning at a time.
The work software could never touch
For forty years, business software was built to record work. Warehouse management systems, ERPs, scheduling tools: clean ledgers of what already happened or was planned. But all of the judgment and coordination happened outside software, and the burden fell to people.
Meanwhile, in the physical world, every leap in how much humanity could produce and afford came from a better way to coordinate physical work: the assembly line and the shipping container, the interstate highway and the barcode. Each one lowered the cost of making and moving things, and living standards climbed for everyone.
But in the past few decades, the gains have slowed. American manufacturing productivity grew about 3.4% a year from the late 1980s through 2007; since 2010 it has actually declined. Construction productivity has been flat for sixty years. Part of the explanation is uncomfortable: we stopped trying. For thirty years, the answer to every hard production problem was to move it somewhere with cheaper labor. Offshoring became a substitute for getting better. While the world of software blossomed, physical productivity atrophied.
Consider what running one building actually takes. One of the best operators we know ran a 750,000-square-foot distribution center where everything was measured: 130 cartons per man-hour coming off a container, 17 pallet putaways an hour, 120 cartons an hour on the pick. He built his staffing model on the cubic volume of inbound cartons, because it told him when the work would spike before the orders did. Years of hand-built spreadsheets kept temp overtime below 1%. Then purchasing consolidated shipments at the port in China and didn’t tell the warehouse. Containers that used to carry two SKUs started arriving with twelve, labor costs jumped double digits in a season, and the model had to be rebuilt by hand, again.
That system had two blind spots. The intelligence that ran that building lived in one person’s spreadsheets. And even at its best, it stopped at the walls: the decision that broke it was made three departments away, by someone correctly optimizing their own number. Operational excellence is still artisanal, built by hand and blind to the rest of the chain. The WMS and the ERP recorded all of it, but didn’t understand it. The real coordination layer of the supply chain, for forty years, has been a spreadsheet and a phone.
For the first time, software can read everything happening across an operation, reason about it, and act. The cost of a well-made decision is collapsing the way the cost of information collapsed a generation ago. The first era of software recorded the work. The second, the one we just entered, can reason about it. The third era is the one we’re building toward: software that orchestrates the work itself, coordinating everyone and everything that gets it done. When those thousands of daily decisions get made with full context, the physical world starts compounding the way the digital world has for thirty years. It’s how the abundance software created in the digital world finally reaches the physical one.
This must happen now
The industrial economy — everything made, moved, and stored — is roughly a third of US GDP, and the ground beneath it is shifting. The workforce that keeps it moving is aging out, and fewer young people are choosing to plug the gap. At the same time, production is coming home. Manufacturing construction spending roughly tripled between 2020 and 2024, with a historic tailwind of data centers to propel it even further in the coming years. Every one of those buildings has to be stocked and staffed. And this time there’s no cheaper somewhere-else to send the hard problems: the productivity has to come from running the operation itself better.
In this future, the people doing this work matter more, not less, and every hour of their effort has to count for more. Getting there takes higher productivity from both people and machines. Robots will carry more of the physical work every year, but a robot is a pair of hands, and hands don’t know what needs to be done at your dock at 4 a.m. What the physical economy needs is a brain above the operation deciding what gets done, by whom or by what, in what order.
2027
We’re building the layer that does the coordinating. We call it Neo. We built an AI agent to run our own operations first: it answers questions, makes phone calls, reviews and updates timesheets, and takes actions across our systems, and our own operators use it every day. Neo is that same agent, sold to the customer. It plugs into the systems a facility already runs: warehouse management, ERP, HRIS, plus the emails and spreadsheets in between, and sees what’s happening across every facility the way no single person can. It takes on the lower-leverage load and decisions — filing carrier claims, chasing a late component across email threads, reconciling detention charges — and keeps people on the tough judgment calls and the messy long tail no machine will touch for a long time. Decisions that used to be reactive become proactive, and Neo learns from every human decision when it needs to escalate. And when the answer is labor, Neo doesn’t stop at a recommendation, because it’s wired into the same marketplace that’s filled a million shifts. “Twelve short on Thursday” becomes twelve vetted workers requested, confirmed, and showing up.
Picture the same facility with Neo, dealing with the consolidated shipments update. The purchasing change surfaces the day the POs change, thirteen weeks before the first twelve-SKU container reaches the dock. Labor is re-planned, crews are trained for the harder breakdown, and costs are repriced in advance. What once took weeks of overtime to absorb becomes a Tuesday. And the supervisor who used to spend hours in a back office rebuilding a staffing model is on the floor, running the system and coaching the team.
This gets much bigger than a single facility. The supply chain is broken in a very specific way: there are few incentives to warn anybody else about what’s coming. A brand runs a promotion, and a regional carrier three layers downstream needs to react, but the signal never arrives in time. One Neo makes one company run better. But when Neo sits inside a 3PL and its supplier and its customer, a disruption in one link becomes a signal every other link acts on in the same afternoon, instead of a surprise that lands on whoever’s standing there when it hits.
Agents across freight, 3PLs, manufacturers, and retailers, talking to each other, a web across the whole chain. Every customer added makes the web smarter for everyone in it. That’s the network effect, and it compounds toward a supply chain that can finally operate on itself.
We earn this the same way we earned our place on the floor: embedding with customers and building solutions, one morning at a time.
The supply chain is the infrastructure everyone depends on and no one thinks about. When it breaks, we all feel it. When it runs well, GDP growth accelerates, the economy bends under pressure instead of snapping, and there’s an industrial base this country can build on again.
That’s the real prize. For a century, the physical world has been the one place software couldn’t reach, and material progress slowed as it stalled there. Bringing intelligence to the floor restarts that engine, and extends to the physical economy the same abundance software already brought to the digital one. The floor is where the modern world gets built. It’s time to bring it online.
