Where AI Is Gaining Traction in Aircraft Maintenance

Maintenance, repair, and overhaul (MRO), or the work of keeping aircraft airworthy, accounts for more than $130B in annual global spend, and is one of the least modernized corners of aviation, with much of it still running on legacy systems and manual workflows. This also makes it one of the most interesting places to build.

Last month, I joined a panel on artificial intelligence in maintenance at MRO Americas in Orlando. The operators there were well past the hype, comparing the tools they already run, where those tools earn their trust, where they fall short, and what it takes to win over a skeptical shop floor. In this piece, I share the SKY VC perspective about where AI is genuinely taking hold in MRO today, and where it heads next.

Where AI Is Starting to Take Hold

In our conversations with airlines and MROs, one shift keeps surfacing: maintenance is moving from a set of discrete, reactive workflows into a predictive, coordinated, data-driven operation. Three themes capture where that's happening, and where operators actually want to go deeper:

• Economic leakage and supply chain intelligence: AI is making fragmented maintenance economics actionable, across warranty recovery, tariffs, inventory, and the reliability gains from predictive maintenance. 

• Maintenance orchestration and collaborative execution: real-time coordination layers that connect the people and systems involved in a single repair.
  
  
• Robotics, inspection, and the digital chain of custody: drones, borescope tools, and the computer vision behind them give operators objective, verifiable readings of an aircraft's condition, and the foundation for a digital twin.
  

That last point, the "digital chain of custody," is where we've seen the most telling shift. The drone-inspection companies and computer-vision tools for engine internals in our network were built to catch defects in images, but the records they generate as a byproduct have become the more valuable output: every finding is timestamped in a format that auditors accept. This trail of verified records matters not only to airlines but also to MROs, aircraft lessors, and other industry bodies. An inspection tool repositioned as a records product reaches a bigger buyer and a bigger market.

Empowering Skilled Workers

The AI we've seen move the needle in MRO is designed to make existing maintenance teams more effective at what they already do, not to replace them. 

The Federal Aviation Administration (FAA) and European Union Aviation Safety Agency (EASA) require a qualified person to sign off on all decisions. Take compliance monitoring: a single aircraft is governed by a web of contracts and regulations, from lease return conditions to service agreements and airworthiness directives. AI can check the work against all of those terms at once, flag what falls short, and pull the supporting documentation together, while a qualified person reviews each exception and owns the call before anything is filed. Parts reconciliation and records review follow the same cadence, with the model doing the legwork and a person ultimately responsible for what goes out.

We are also hearing from operators who want to use AI to train and scale their teams faster. Maintenance is chronically short-staffed, and turnover is high, so getting new technicians up to speed sooner has become a real priority.

Build vs. Buy

The build-versus-buy question comes up in every MRO conversation we have. Take an example we’ve seen firsthand: one major US airline chose to build its own warranty management tool. Another competitive carrier bought from a vendor rather than building internally. Both reached a working solution, and the airline that bought is now saving $25M+ a year.

For most operators, the decision comes down to speed to value, how much ongoing engineering they want their own teams to carry, and whether warranty recovery is a capability worth owning or a problem better handed to a focused vendor. An airline with its own MRO has a real case for building, given how mission-critical the work is and how much maintenance data it already holds in-house. The rest tend to get there faster with a vendor, and a focused one builds an edge no single airline can match: it learns from every operator it serves, not just one. The wedge for founders is owning a single workflow end-to-end and producing savings that an airline's finance team can quote with confidence.

Coordination Comes Next

So far, most AI applications in MRO have been point solutions inside individual workflows. The bigger opportunity we see lies in the layer that connects those workflows.

Maintenance runs on coordination among many parties. A single grounded aircraft pulls in the airline's operations team, the mechanics on the line, the parts team, the third-party MRO, the original equipment manufacturer's technical rep, and the regulatory parties behind it all. Much of that still happens over phone calls, group chats, and systems that don't talk to each other, and the cost shows up in ground hours and inflated costs that an airline never gets back.

We’ve already seen several companies trying to build that coordination layer for maintenance – capturing mechanics’ conversations on the shop floor as structured data, pushing real-time status across all parties, and feeding cleaner records back into the systems in place.

The longer-term vision is the aircraft digital twin: a single, complete digital record of an aircraft, tied to its manufacturer's serial number. It would hold the aircraft's condition, maintenance history, and operational data in one place for all stakeholders. Luckily, the building blocks already exist. Computer vision handles documentation, natural language processing expedites analysis, and predictive models harmonize scheduling. The missing piece is that they are usually built in isolation. A real twin depends on those tools sharing data; five systems that cannot talk to each other are just five disconnected dashboards. Connecting them from the start is the harder and more valuable work.

What We're Watching

MRO is a hard market, and its buyers are conservative for valid reasons. Sales cycles are long, and reference customers move slowly. For founders willing to work through that, the reward is a durable position in a market that rarely switches vendors.

We're spending our time across all three themes, and we'd love to talk to founders building against any of them. Reach out at inbound@sky-vc.com.