Autonoma: Autonomy Where Failure Isn’t an Option

Autonomy is arriving in aviation—but not as quickly as its advocates once promised.

Reporting from The Wall Street Journal and Financial Times has documented the pressure carriers face: volatile weather, labor constraints, aging infrastructure, and razor-thin margins, all while customers demand reliability. The industry’s response has been a wave of AI-driven optimization tools meant to improve scheduling, routing, and ground operations. What those tools often lack is a way to be tested safely.

That gap is where Valor portfolio company Autonoma is redefining operational edge with AI accountability.

Over the past year, the Auburn-based startup, led by founder and CEO Will Bryan, has sharpened its focus on aviation—deploying its digital simulation platform inside real airline and airport environments and extending similar work into defense. Rather than selling autonomy as a black box, Autonoma has positioned itself as a validation layer: a place where AI-driven decisions can be tested, stressed, and compared against reality before they are deployed live.

“We create a virtual environment where teams can test decisions before they ever touch the real world,” Bryan said during a recent interview led by Autonoma board director and Valor operating partner Jean-Luc VanHulst. “It looks like a video game, but it’s grounded in real physics, weather, geometry, and operational constraints.”

Autonoma’s platform—called the AutoVerse—is a high-fidelity, three-dimensional simulation of complex environments. It ingests both historical and live data from sensors, aircraft, and ground vehicles, allowing operators to run hundreds or thousands of scenarios and compare outcomes before committing to a plan.

That capability has proven particularly relevant in aviation, where most delays originate on the ground rather than in the air. Gates, crews, service vehicles, and external airline interactions all contribute to cascading effects that traditional planning tools struggle to model.

Bryan offered a concrete example drawn from Autonoma’s work with Delta.

“Even the most on-time airlines are only around sixty percent,” he said. “When a plane arrives late, it often has to wait for a gate. That delay ripples—every flight behind it gets affected.” Delta has been developing AI-based gating and scheduling models to manage those dynamics. But before applying new logic inside a live airport like Atlanta—where a misstep could affect thousands of flights in a single day—the airline needed a way to validate those decisions.

Autonoma built a full simulation of the airport environment, integrating Delta’s live and historical data.

Inside the AutoVerse, Delta can run proposed schedules repeatedly, identify bottlenecks, and compare outcomes before the day begins.

“They can run through hundreds or thousands of simulations,” Bryan said. “Then choose the option that has the least impact on the customer, the best throughput, and the fewest downstream issues.”

During the session, VanHulst paused the demo to clarify what the room was seeing. “So what we’re looking at here,” he asked, “this isn’t drone footage. This isn’t real?”

“No,” Bryan replied. “This is entirely our software.”

That distinction—between visualization and validation—has defined Autonoma’s progress over the last year. The company has moved through phased deployments: building simulation models, integrating live data feeds, comparing simulated outcomes to real-world results, and embedding directly into internal airline systems.

“For us to say the simulation is accurate, we have to prove it,” Bryan said. “We run a full schedule, compare it to what really happened, and keep tightening that variance.”

Discipline shapes Autonoma’s strategic direction.

While the underlying technology could extend into adjacent domains, the company has chosen to focus deeply on aviation before expanding further. In parallel, it has begun applying similar approaches with the Air Force and Navy, where validation and consequence are equally central.

“Right now, we’re very focused on aviation,” Bryan said. “Ask me in another year about expanding more broadly.”

Industry analysts have increasingly emphasized this phase of AI adoption. Research from firms such as McKinsey and Gartner over the past two years points to a shift away from experimentation toward governed, validated deployment—particularly in regulated and safety-critical environments.

Autonoma’s work reflects that transition: autonomy as infrastructure.

VanHulst underscored that perspective during the discussion, noting that the hardest part of autonomy isn’t the model itself, but making sure it behaves predictably when reality intervenes.