They didn’t want a smarter dispatcher. They wanted a bigger market.

At Airspace, exploring expansion opportunities led us toward small courier operators, the independent delivery businesses the logistics industry calls “agents.” Early surveys and interviews were encouraging, and there looked to be a real chance to port the transportation-management system we had been dogfooding for years into a new market.

The smartest layer inside it was routing and dispatch intelligence, the software that works out how a time-critical shipment should move. Putting that in the hands of operators running their own small fleets looked like the obvious offer.

I led the research to test whether the opportunity was as promising as it looked. We ran it as a weekly design sprint: write down an assumption plainly enough that an operator could contradict it, build the smallest honest version of it, and put it in front of the people who ran these businesses. The surprises started immediately.

They already had the inside of the operation handled.

Smart Dispatch was our headline capability, the software that picks the right driver for each job. Auto Delay Detection sat just behind it. In the interviews, neither registered as an urgent problem. Most operators worked a territory small enough to hold in their heads, and a text thread with their drivers did the job well enough that a better answer was not worth learning a new system for.

Some of them went further than indifference. A few of the smaller operators told us plainly that dispatch automation would be a drawback: a system that assigned the work would weaken their relationships with their own drivers. Operators felt the personal touch was key to keeping their drivers satisfied.

The part they could not solve was the edge.

What operators kept returning to was the boundary between their business and the market around it. They wanted a way to signal that they had capacity and have work show up, which happened to be something Airspace was in an unusual position to supply. They wanted to raise their prices when they were overwhelmed, rather than quoting the same familiar number into a week they could not cover.

Everything inside the operation flexed. Drivers came and went, volume rose and fell, and a good week looked nothing like a thin one. Their finances were static. One rate card, one set of terms, and no way to answer either kind of pressure.

The rest of what they asked for pointed the same direction. Instant quotes. Order-entry forms they could hand to a customer. Invoicing customers, and paying drivers. A record of which drivers preferred which work and when they were available, so the operator could make a better call. Not so the software could make it for them.

A valve, not a smarter algorithm.

What came out of the research was narrower than a port of our dispatch or delay intelligence, and more useful. Take the fundamentals of the transportation-management system we already ran, and connect an operator’s capacity to the market in both directions. When their own pipeline was dry, work could reach them through our courier network. When they were slammed, their pricing could move. Automation when they needed it, and more importantly, when they asked for it.

With a better understanding of what kind of platform it would take to capture the agent market we decided it wasn’t our top priority. Years later, a different kind of courier operation came onto our radar: companies that wanted to route fleets of their own, with our contractors available when their volume overflowed. We were ready. Fleet Manager carries the lessons forward: lead with the fundamentals, and allow operators to opt into the intelligence when they are ready.