The airfield business case, and everything in between

Welcome to the first edition of Real-World Autonomy. Each month we work through what it takes to make autonomous systems real, from the operations to the economics to the timing. Less hype, more signal.

This month, the airfield. The business case for automating the ground fleet, how integration actually works, and a week from the ramp at the world's busiest airport. The technology is ready. The gap is everything in between.

The number nobody will say out loud

Everyone agrees airside fleets will go autonomous. Almost nobody will say when it pays.

So we modeled it at the scale of a major hub. A thousand manual tugs becomes a shared pool of about 400. The capital expenditure premium is real. So is the $149 million in annual savings, eventually.

The catch is timing. The savings are not available on day one, and the phase where you are running both fleets is the phase that kills programs that did not plan for it.

Integrating autonomy into aviation operations

The business case is one thing. The method is another.

The work is not the vision, it is the integration: where programs stall, what real readiness looks like, and the steps that move an operator from curiosity to a system that runs every day. If the numbers convince you the timing is coming, that is where preparation starts.

Life of an AV integrator: one week at Hartsfield-Jackson

I spent a week at Hartsfield-Jackson Atlanta International Airport, the world's busiest airport. This is what the job actually looks like for an autonomous vehicle integrator on the ground.

It starts the way every deployment really starts, with a commute. From there it moves through an all-day risk assessment for an automated baggage tug, a running reminder to never lose sight of the regulatory framework, a pilot study for an autonomous wheelchair where the cargo is people and the safety bar climbs, and a deep-dive session on autonomous FOD detection and removal, because a single loose bolt on a runway is a very expensive problem.

Three different systems, all advancing at once, each with its own stakes. Real-world autonomy is built one careful, grounded day at a time.

Before you go

If your team is wrestling with any of these systems, the hardest part is rarely the hardware. It is the integration layer.

This edition also appears on LinkedIn. Read and subscribe there.